mdw@git.distorted.org.uk Git - disorder/blob - lib/mime.c

   1 /*
   2  * This file is part of DisOrder
   3  * Copyright (C) 2005, 2007 Richard Kettlewell
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful, but
  11  * WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13  * General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License
  16  * along with this program; if not, write to the Free Software
  17  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
  18  * USA
  19  */
  20 /** @file lib/mime.c
  21  * @brief Support for MIME and allied protocols
  22  */
  23
  24 #include <config.h>
  25 #include "types.h"
  26
  27 #include <string.h>
  28 #include <ctype.h>
  29
  30 #include <stdio.h>
  31
  32 #include "mem.h"
  33 #include "mime.h"
  34 #include "vector.h"
  35 #include "hex.h"
  36 #include "log.h"
  37 #include "base64.h"
  38
  39 /** @brief Match whitespace characters */
  40 static int whitespace(int c) {
  41   switch(c) {
  42   case ' ':
  43   case '\t':
  44   case '\r':
  45   case '\n':
  46     return 1;
  47   default:
  48     return 0;
  49   }
  50 }
  51
  52 /** @brief Match RFC2045 tspecial characters */
  53 static int tspecial(int c) {
  54   switch(c) {
  55   case '(':
  56   case ')':
  57   case '<':
  58   case '>':
  59   case '@':
  60   case ',':
  61   case ';':
  62   case ':':
  63   case '\\':
  64   case '"':
  65   case '/':
  66   case '[':
  67   case ']':
  68   case '?':
  69   case '=':
  70     return 1;
  71   default:
  72     return 0;
  73   }
  74 }
  75
  76 /** @brief Match RFC2616 seprator characters */
  77 static int http_separator(int c) {
  78   switch(c) {
  79   case '(':
  80   case ')':
  81   case '<':
  82   case '>':
  83   case '@':
  84   case ',':
  85   case ';':
  86   case ':':
  87   case '\\':
  88   case '"':
  89   case '/':
  90   case '[':
  91   case ']':
  92   case '?':
  93   case '=':
  94   case '{':
  95   case '}':
  96   case ' ':
  97   case '\t':
  98     return 1;
  99   default:
 100     return 0;
 101   }
 102 }
 103
 104 /** @brief Match CRLF */
 105 static int iscrlf(const char *ptr) {
 106   return ptr[0] == '\r' && ptr[1] == '\n';
 107 }
 108
 109 /** @brief Skip whitespace
 110  * @param s Pointer into string
 111  * @param rfc822_comments If true, skip RFC822 nested comments
 112  * @return Pointer into string after whitespace
 113  */
 114 static const char *skipwhite(const char *s, int rfc822_comments) {
 115   int c, depth;
 116
 117   for(;;) {
 118     switch(c = *s) {
 119     case ' ':
 120     case '\t':
 121     case '\r':
 122     case '\n':
 123       ++s;
 124       break;
 125     case '(':
 126       if(!rfc822_comments)
 127         return s;
 128       ++s;
 129       depth = 1;
 130       while(*s && depth) {
 131         c = *s++;
 132         switch(c) {
 133         case '(': ++depth; break;
 134         case ')': --depth; break;
 135         case '\\':
 136           if(!*s) return 0;
 137           ++s;
 138           break;
 139         }
 140       }
 141       if(depth) return 0;
 142       break;
 143     default:
 144       return s;
 145     }
 146   }
 147 }
 148
 149 /** @brief Test for a word character
 150  * @param c Character to test
 151  * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
 152  * @return 1 if @p c is a word character, else 0
 153  */
 154 static int iswordchar(int c, int (*special)(int)) {
 155   return !(c <= ' ' || c > '~' || special(c));
 156 }
 157
 158 /** @brief Parse an RFC1521/RFC2616 word
 159  * @param s Pointer to start of word
 160  * @param valuep Where to store value
 161  * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
 162  * @return Pointer just after end of word or NULL if there's no word
 163  *
 164  * A word is a token or a quoted-string.
 165  */
 166 static const char *parseword(const char *s, char **valuep,
 167                              int (*special)(int)) {
 168   struct dynstr value[1];
 169   int c;
 170
 171   dynstr_init(value);
 172   if(*s == '"') {
 173     ++s;
 174     while((c = *s++) != '"') {
 175       switch(c) {
 176       case '\\':
 177         if(!(c = *s++)) return 0;
 178       default:
 179         dynstr_append(value, c);
 180         break;
 181       }
 182     }
 183     if(!c) return 0;
 184   } else {
 185     if(!iswordchar((unsigned char)*s, special))
 186       return NULL;
 187     dynstr_init(value);
 188     while(iswordchar((unsigned char)*s, special))
 189       dynstr_append(value, *s++);
 190   }
 191   dynstr_terminate(value);
 192   *valuep = value->vec;
 193   return s;
 194 }
 195
 196 /** @brief Parse an RFC1521/RFC2616 token
 197  * @param s Pointer to start of token
 198  * @param valuep Where to store value
 199  * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
 200  * @return Pointer just after end of token or NULL if there's no token
 201  */
 202 static const char *parsetoken(const char *s, char **valuep,
 203                               int (*special)(int)) {
 204   if(*s == '"') return 0;
 205   return parseword(s, valuep, special);
 206 }
 207
 208 /** @brief Parse a MIME content-type field
 209  * @param s Start of field
 210  * @param typep Where to store type
 211  * @param parameternamep Where to store parameter name
 212  * @param parametervaluep Wher to store parameter value
 213  * @return 0 on success, non-0 on error
 214  *
 215  * See <a href="http://tools.ietf.org/html/rfc2045#section-5">RFC 2045 s5</a>.
 216  */
 217 int mime_content_type(const char *s,
 218                       char **typep,
 219                       char **parameternamep,
 220                       char **parametervaluep) {
 221   struct dynstr type, parametername;
 222
 223   dynstr_init(&type);
 224   if(!(s = skipwhite(s, 1))) return -1;
 225   if(!*s) return -1;
 226   while(*s && !tspecial(*s) && !whitespace(*s))
 227     dynstr_append(&type, tolower((unsigned char)*s++));
 228   if(!(s = skipwhite(s, 1))) return -1;
 229   if(*s++ != '/') return -1;
 230   dynstr_append(&type, '/');
 231   if(!(s = skipwhite(s, 1))) return -1;
 232   while(*s && !tspecial(*s) && !whitespace(*s))
 233     dynstr_append(&type, tolower((unsigned char)*s++));
 234   if(!(s = skipwhite(s, 1))) return -1;
 235
 236   if(*s == ';') {
 237     dynstr_init(&parametername);
 238     ++s;
 239     if(!(s = skipwhite(s, 1))) return -1;
 240     if(!*s) return -1;
 241     while(*s && !tspecial(*s) && !whitespace(*s))
 242       dynstr_append(&parametername, tolower((unsigned char)*s++));
 243     if(!(s = skipwhite(s, 1))) return -1;
 244     if(*s++ != '=') return -1;
 245     if(!(s = skipwhite(s, 1))) return -1;
 246     if(!(s = parseword(s, parametervaluep, tspecial))) return -1;
 247     if(!(s = skipwhite(s, 1))) return -1;
 248     dynstr_terminate(&parametername);
 249     *parameternamep = parametername.vec;
 250   } else
 251     *parametervaluep = *parameternamep = 0;
 252   dynstr_terminate(&type);
 253   *typep = type.vec;
 254   return 0;
 255 }
 256
 257 /** @brief Parse a MIME message
 258  * @param s Start of message
 259  * @param callback Called for each header field
 260  * @param u Passed to callback
 261  * @return Pointer to decoded body (might be in original string), or NULL on error
 262  *
 263  * This does an RFC 822-style parse and honors Content-Transfer-Encoding as
 264  * described in <a href="http://tools.ietf.org/html/rfc2045#section-6">RFC 2045
 265  * s6</a>.  @p callback is called for each header field encountered, in order,
 266  * with ASCII characters in the header name forced to lower case.
 267  */
 268 const char *mime_parse(const char *s,
 269                        int (*callback)(const char *name, const char *value,
 270                                        void *u),
 271                        void *u) {
 272   struct dynstr name, value;
 273   char *cte = 0, *p;
 274
 275   while(*s && !iscrlf(s)) {
 276     dynstr_init(&name);
 277     dynstr_init(&value);
 278     while(*s && !tspecial(*s) && !whitespace(*s))
 279       dynstr_append(&name, tolower((unsigned char)*s++));
 280     if(!(s = skipwhite(s, 1))) return 0;
 281     if(*s != ':') return 0;
 282     ++s;
 283     while(*s && !(*s == '\n' && !(s[1] == ' ' || s[1] == '\t')))
 284       dynstr_append(&value, *s++);
 285     if(*s) ++s;
 286     dynstr_terminate(&name);
 287     dynstr_terminate(&value);
 288     if(!strcmp(name.vec, "content-transfer-encoding")) {
 289       cte = xstrdup(value.vec);
 290       for(p = cte; *p; p++)
 291         *p = tolower((unsigned char)*p);
 292     }
 293     if(callback(name.vec, value.vec, u)) return 0;
 294   }
 295   if(*s) s += 2;
 296   if(cte) {
 297     if(!strcmp(cte, "base64")) return mime_base64(s, 0);
 298     if(!strcmp(cte, "quoted-printable")) return mime_qp(s);
 299   }
 300   return s;
 301 }
 302
 303 /** @brief Match the boundary string */
 304 static int isboundary(const char *ptr, const char *boundary, size_t bl) {
 305   return (ptr[0] == '-'
 306           && ptr[1] == '-'
 307           && !strncmp(ptr + 2, boundary, bl)
 308           && (iscrlf(ptr + bl + 2)
 309               || (ptr[bl + 2] == '-'
 310                   && ptr[bl + 3] == '-'
 311                   && (iscrlf(ptr + bl + 4) || *(ptr + bl + 4) == 0))));
 312 }
 313
 314 /** @brief Match the final boundary string */
 315 static int isfinal(const char *ptr, const char *boundary, size_t bl) {
 316   return (ptr[0] == '-'
 317           && ptr[1] == '-'
 318           && !strncmp(ptr + 2, boundary, bl)
 319           && ptr[bl + 2] == '-'
 320           && ptr[bl + 3] == '-'
 321           && (iscrlf(ptr + bl + 4) || *(ptr + bl + 4) == 0));
 322 }
 323
 324 /** @brief Parse a multipart MIME body
 325  * @param s Start of message
 326  * @param callback Callback for each part
 327  * @param boundary Boundary string
 328  * @param u Passed to callback
 329  * @return 0 on success, non-0 on error
 330  *
 331  * See <a href="http://tools.ietf.org/html/rfc2046#section-5.1">RFC 2046
 332  * s5.1</a>.  @p callback is called for each part (not yet decoded in any way)
 333  * in succession; you should probably call mime_parse() for each part.
 334  */
 335 int mime_multipart(const char *s,
 336                    int (*callback)(const char *s, void *u),
 337                    const char *boundary,
 338                    void *u) {
 339   size_t bl = strlen(boundary);
 340   const char *start, *e;
 341   int ret;
 342
 343   /* We must start with a boundary string */
 344   if(!isboundary(s, boundary, bl))
 345     return -1;
 346   /* Keep going until we hit a final boundary */
 347   while(!isfinal(s, boundary, bl)) {
 348     s = strstr(s, "\r\n") + 2;
 349     start = s;
 350     while(!isboundary(s, boundary, bl)) {
 351       if(!(e = strstr(s, "\r\n")))
 352         return -1;
 353       s = e + 2;
 354     }
 355     if((ret = callback(xstrndup(start,
 356                                 s == start ? 0 : s - start - 2),
 357                        u)))
 358       return ret;
 359   }
 360   return 0;
 361 }
 362
 363 /** @brief Parse an RFC2388-style content-disposition field
 364  * @param s Start of field
 365  * @param dispositionp Where to store disposition
 366  * @param parameternamep Where to store parameter name
 367  * @param parametervaluep Wher to store parameter value
 368  * @return 0 on success, non-0 on error
 369  *
 370  * See <a href="http://tools.ietf.org/html/rfc2388#section-3">RFC 2388 s3</a>
 371  * and <a href="http://tools.ietf.org/html/rfc2183">RFC 2183</a>.
 372  */
 373 int mime_rfc2388_content_disposition(const char *s,
 374                                      char **dispositionp,
 375                                      char **parameternamep,
 376                                      char **parametervaluep) {
 377   struct dynstr disposition, parametername;
 378
 379   dynstr_init(&disposition);
 380   if(!(s = skipwhite(s, 1))) return -1;
 381   if(!*s) return -1;
 382   while(*s && !tspecial(*s) && !whitespace(*s))
 383     dynstr_append(&disposition, tolower((unsigned char)*s++));
 384   if(!(s = skipwhite(s, 1))) return -1;
 385
 386   if(*s == ';') {
 387     dynstr_init(&parametername);
 388     ++s;
 389     if(!(s = skipwhite(s, 1))) return -1;
 390     if(!*s) return -1;
 391     while(*s && !tspecial(*s) && !whitespace(*s))
 392       dynstr_append(&parametername, tolower((unsigned char)*s++));
 393     if(!(s = skipwhite(s, 1))) return -1;
 394     if(*s++ != '=') return -1;
 395     if(!(s = skipwhite(s, 1))) return -1;
 396     if(!(s = parseword(s, parametervaluep, tspecial))) return -1;
 397     if(!(s = skipwhite(s, 1))) return -1;
 398     dynstr_terminate(&parametername);
 399     *parameternamep = parametername.vec;
 400   } else
 401     *parametervaluep = *parameternamep = 0;
 402   dynstr_terminate(&disposition);
 403   *dispositionp = disposition.vec;
 404   return 0;
 405 }
 406
 407 /** @brief Convert MIME quoted-printable
 408  * @param s Quoted-printable data
 409  * @return Decoded data
 410  *
 411  * See <a href="http://tools.ietf.org/html/rfc2045#section-6.7">RFC 2045
 412  * s6.7</a>.
 413  */
 414 char *mime_qp(const char *s) {
 415   struct dynstr d;
 416   int c, a, b;
 417   const char *t;
 418
 419   dynstr_init(&d);
 420   while((c = *s++)) {
 421     switch(c) {
 422     case '=':
 423       if((a = unhexdigitq(s[0])) != -1
 424          && (b = unhexdigitq(s[1])) != -1) {
 425         dynstr_append(&d, a * 16 + b);
 426         s += 2;
 427       } else {
 428         t = s;
 429         while(*t == ' ' || *t == '\t') ++t;
 430         if(iscrlf(t)) {
 431           /* soft line break */
 432           s = t + 2;
 433         } else
 434           return 0;
 435       }
 436       break;
 437     case ' ':
 438     case '\t':
 439       t = s;
 440       while(*t == ' ' || *t == '\t') ++t;
 441       if(iscrlf(t))
 442         /* trailing space is always eliminated */
 443         s = t;
 444       else
 445         dynstr_append(&d, c);
 446       break;
 447     default:
 448       dynstr_append(&d, c);
 449       break;
 450     }
 451   }
 452   dynstr_terminate(&d);
 453   return d.vec;
 454 }
 455
 456 /** @brief Parse a RFC2109 Cookie: header
 457  * @param s Header field value
 458  * @param cd Where to store result
 459  * @return 0 on success, non-0 on error
 460  *
 461  * See <a href="http://tools.ietf.org/html/rfc2109">RFC 2109</a>.
 462  */
 463 int parse_cookie(const char *s,
 464                  struct cookiedata *cd) {
 465   char *n = 0, *v = 0;
 466
 467   memset(cd, 0, sizeof *cd);
 468   s = skipwhite(s, 0);
 469   while(*s) {
 470     /* Skip separators */
 471     if(*s == ';' || *s == ',') {
 472       ++s;
 473       s = skipwhite(s, 0);
 474       continue;
 475     }
 476     if(!(s = parsetoken(s, &n, http_separator))) return -1;
 477     s = skipwhite(s, 0);
 478     if(*s++ != '=') return -1;
 479     s = skipwhite(s, 0);
 480     if(!(s = parseword(s, &v, http_separator))) return -1;
 481     if(n[0] == '$') {
 482       /* Some bit of meta-information */
 483       if(!strcmp(n, "$Version"))
 484         cd->version = v;
 485       else if(!strcmp(n, "$Path")) {
 486         if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].path == 0)
 487           cd->cookies[cd->ncookies-1].path = v;
 488         else {
 489           error(0, "redundant $Path in Cookie: header");
 490           return -1;
 491         }
 492       } else if(!strcmp(n, "$Domain")) {
 493         if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].domain == 0)
 494           cd->cookies[cd->ncookies-1].domain = v;
 495         else {
 496           error(0, "redundant $Domain in Cookie: header");
 497           return -1;
 498         }
 499       }
 500     } else {
 501       /* It's a new cookie */
 502       cd->cookies = xrealloc(cd->cookies,
 503                              (cd->ncookies + 1) * sizeof (struct cookie));
 504       cd->cookies[cd->ncookies].name = n;
 505       cd->cookies[cd->ncookies].value = v;
 506       cd->cookies[cd->ncookies].path = 0;
 507       cd->cookies[cd->ncookies].domain = 0;
 508       ++cd->ncookies;
 509     }
 510     s = skipwhite(s, 0);
 511     if(*s && (*s != ',' && *s != ';')) {
 512       error(0, "missing separator in Cookie: header");
 513       return -1;
 514     }
 515   }
 516   return 0;
 517 }
 518
 519 /** @brief Find a named cookie
 520  * @param cd Parse cookie data
 521  * @param name Name of cookie
 522  * @return Cookie structure or NULL if not found
 523  */
 524 const struct cookie *find_cookie(const struct cookiedata *cd,
 525                                  const char *name) {
 526   int n;
 527
 528   for(n = 0; n < cd->ncookies; ++n)
 529     if(!strcmp(cd->cookies[n].name, name))
 530       return &cd->cookies[n];
 531   return 0;
 532 }
 533
 534 /** @brief RFC822 quoting
 535  * @param s String to quote
 536  * @param force If non-0, always quote
 537  * @return Possibly quoted string
 538  */
 539 char *quote822(const char *s, int force) {
 540   const char *t;
 541   struct dynstr d[1];
 542   int c;
 543
 544   if(!force) {
 545     /* See if we need to quote */
 546     for(t = s; (c = (unsigned char)*t); ++t) {
 547       if(tspecial(c) || http_separator(c) || whitespace(c))
 548         break;
 549     }
 550     if(*t)
 551       force = 1;
 552   }
 553
 554   if(!force)
 555     return xstrdup(s);
 556
 557   dynstr_init(d);
 558   dynstr_append(d, '"');
 559   for(t = s; (c = (unsigned char)*t); ++t) {
 560     if(c == '"' || c == '\\')
 561       dynstr_append(d, '\\');
 562     dynstr_append(d, c);
 563   }
 564   dynstr_append(d, '"');
 565   dynstr_terminate(d);
 566   return d->vec;
 567 }
 568
 569 /*
 570 Local Variables:
 571 c-basic-offset:2
 572 comment-column:40
 573 fill-column:79
 574 End:
 575 */