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 #include "kvp.h"
  39
  40 /** @brief Match whitespace characters */
  41 static int whitespace(int c) {
  42   switch(c) {
  43   case ' ':
  44   case '\t':
  45   case '\r':
  46   case '\n':
  47     return 1;
  48   default:
  49     return 0;
  50   }
  51 }
  52
  53 /** @brief Match RFC2045 tspecial characters */
  54 static int tspecial(int c) {
  55   switch(c) {
  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   case '=':
  71     return 1;
  72   default:
  73     return 0;
  74   }
  75 }
  76
  77 /** @brief Match RFC2616 separator characters */
  78 static int http_separator(int c) {
  79   switch(c) {
  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 ' ':
  98   case '\t':
  99     return 1;
 100   default:
 101     return 0;
 102   }
 103 }
 104
 105 /** @brief Match CRLF */
 106 static int iscrlf(const char *ptr) {
 107   return ptr[0] == '\r' && ptr[1] == '\n';
 108 }
 109
 110 /** @brief Skip whitespace
 111  * @param s Pointer into string
 112  * @param rfc822_comments If true, skip RFC822 nested comments
 113  * @return Pointer into string after whitespace
 114  */
 115 static const char *skipwhite(const char *s, int rfc822_comments) {
 116   int c, depth;
 117
 118   for(;;) {
 119     switch(c = *s) {
 120     case ' ':
 121     case '\t':
 122     case '\r':
 123     case '\n':
 124       ++s;
 125       break;
 126     case '(':
 127       if(!rfc822_comments)
 128         return s;
 129       ++s;
 130       depth = 1;
 131       while(*s && depth) {
 132         c = *s++;
 133         switch(c) {
 134         case '(': ++depth; break;
 135         case ')': --depth; break;
 136         case '\\':
 137           if(!*s) return 0;
 138           ++s;
 139           break;
 140         }
 141       }
 142       if(depth) return 0;
 143       break;
 144     default:
 145       return s;
 146     }
 147   }
 148 }
 149
 150 /** @brief Test for a word character
 151  * @param c Character to test
 152  * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
 153  * @return 1 if @p c is a word character, else 0
 154  */
 155 static int iswordchar(int c, int (*special)(int)) {
 156   return !(c <= ' ' || c > '~' || special(c));
 157 }
 158
 159 /** @brief Parse an RFC1521/RFC2616 word
 160  * @param s Pointer to start of word
 161  * @param valuep Where to store value
 162  * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
 163  * @return Pointer just after end of word or NULL if there's no word
 164  *
 165  * A word is a token or a quoted-string.
 166  */
 167 static const char *parseword(const char *s, char **valuep,
 168                              int (*special)(int)) {
 169   struct dynstr value[1];
 170   int c;
 171
 172   dynstr_init(value);
 173   if(*s == '"') {
 174     ++s;
 175     while((c = *s++) != '"') {
 176       switch(c) {
 177       case '\\':
 178         if(!(c = *s++)) return 0;
 179       default:
 180         dynstr_append(value, c);
 181         break;
 182       }
 183     }
 184     if(!c) return 0;
 185   } else {
 186     if(!iswordchar((unsigned char)*s, special))
 187       return NULL;
 188     dynstr_init(value);
 189     while(iswordchar((unsigned char)*s, special))
 190       dynstr_append(value, *s++);
 191   }
 192   dynstr_terminate(value);
 193   *valuep = value->vec;
 194   return s;
 195 }
 196
 197 /** @brief Parse an RFC1521/RFC2616 token
 198  * @param s Pointer to start of token
 199  * @param valuep Where to store value
 200  * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
 201  * @return Pointer just after end of token or NULL if there's no token
 202  */
 203 static const char *parsetoken(const char *s, char **valuep,
 204                               int (*special)(int)) {
 205   if(*s == '"') return 0;
 206   return parseword(s, valuep, special);
 207 }
 208
 209 /** @brief Parse a MIME content-type field
 210  * @param s Start of field
 211  * @param typep Where to store type
 212  * @param parametersp Where to store parameter list
 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                       struct kvp **parametersp) {
 220   struct dynstr type, parametername;
 221   struct kvp *parameters = 0;
 222   char *parametervalue;
 223
 224   dynstr_init(&type);
 225   if(!(s = skipwhite(s, 1))) return -1;
 226   if(!*s) return -1;
 227   while(*s && !tspecial(*s) && !whitespace(*s))
 228     dynstr_append(&type, tolower((unsigned char)*s++));
 229   if(!(s = skipwhite(s, 1))) return -1;
 230   if(*s++ != '/') return -1;
 231   dynstr_append(&type, '/');
 232   if(!(s = skipwhite(s, 1))) return -1;
 233   while(*s && !tspecial(*s) && !whitespace(*s))
 234     dynstr_append(&type, tolower((unsigned char)*s++));
 235   if(!(s = skipwhite(s, 1))) return -1;
 236
 237   while(*s == ';') {
 238     dynstr_init(&parametername);
 239     ++s;
 240     if(!(s = skipwhite(s, 1))) return -1;
 241     if(!*s) return -1;
 242     while(*s && !tspecial(*s) && !whitespace(*s))
 243       dynstr_append(&parametername, tolower((unsigned char)*s++));
 244     if(!(s = skipwhite(s, 1))) return -1;
 245     if(*s++ != '=') return -1;
 246     if(!(s = skipwhite(s, 1))) return -1;
 247     if(!(s = parseword(s, &parametervalue, tspecial))) return -1;
 248     if(!(s = skipwhite(s, 1))) return -1;
 249     dynstr_terminate(&parametername);
 250     kvp_set(&parameters, parametername.vec, parametervalue);
 251   }
 252   dynstr_terminate(&type);
 253   *typep = type.vec;
 254   *parametersp = parameters;
 255   return 0;
 256 }
 257
 258 /** @brief Parse a MIME message
 259  * @param s Start of message
 260  * @param callback Called for each header field
 261  * @param u Passed to callback
 262  * @return Pointer to decoded body (might be in original string), or NULL on error
 263  *
 264  * This does an RFC 822-style parse and honors Content-Transfer-Encoding as
 265  * described in <a href="http://tools.ietf.org/html/rfc2045#section-6">RFC 2045
 266  * s6</a>.  @p callback is called for each header field encountered, in order,
 267  * with ASCII characters in the header name forced to lower case.
 268  */
 269 const char *mime_parse(const char *s,
 270                        int (*callback)(const char *name, const char *value,
 271                                        void *u),
 272                        void *u) {
 273   struct dynstr name, value;
 274   char *cte = 0, *p;
 275
 276   while(*s && !iscrlf(s)) {
 277     dynstr_init(&name);
 278     dynstr_init(&value);
 279     while(*s && !tspecial(*s) && !whitespace(*s))
 280       dynstr_append(&name, tolower((unsigned char)*s++));
 281     if(!(s = skipwhite(s, 1))) return 0;
 282     if(*s != ':') return 0;
 283     ++s;
 284     while(*s && !(*s == '\n' && !(s[1] == ' ' || s[1] == '\t')))
 285       dynstr_append(&value, *s++);
 286     if(*s) ++s;
 287     dynstr_terminate(&name);
 288     dynstr_terminate(&value);
 289     if(!strcmp(name.vec, "content-transfer-encoding")) {
 290       cte = xstrdup(value.vec);
 291       for(p = cte; *p; p++)
 292         *p = tolower((unsigned char)*p);
 293     }
 294     if(callback(name.vec, value.vec, u)) return 0;
 295   }
 296   if(*s) s += 2;
 297   if(cte) {
 298     if(!strcmp(cte, "base64")) return mime_base64(s, 0);
 299     if(!strcmp(cte, "quoted-printable")) return mime_qp(s);
 300   }
 301   return s;
 302 }
 303
 304 /** @brief Match the boundary string */
 305 static int isboundary(const char *ptr, const char *boundary, size_t bl) {
 306   return (ptr[0] == '-'
 307           && ptr[1] == '-'
 308           && !strncmp(ptr + 2, boundary, bl)
 309           && (iscrlf(ptr + bl + 2)
 310               || (ptr[bl + 2] == '-'
 311                   && ptr[bl + 3] == '-'
 312                   && (iscrlf(ptr + bl + 4) || *(ptr + bl + 4) == 0))));
 313 }
 314
 315 /** @brief Match the final boundary string */
 316 static int isfinal(const char *ptr, const char *boundary, size_t bl) {
 317   return (ptr[0] == '-'
 318           && ptr[1] == '-'
 319           && !strncmp(ptr + 2, boundary, bl)
 320           && ptr[bl + 2] == '-'
 321           && ptr[bl + 3] == '-'
 322           && (iscrlf(ptr + bl + 4) || *(ptr + bl + 4) == 0));
 323 }
 324
 325 /** @brief Parse a multipart MIME body
 326  * @param s Start of message
 327  * @param callback Callback for each part
 328  * @param boundary Boundary string
 329  * @param u Passed to callback
 330  * @return 0 on success, non-0 on error
 331  *
 332  * See <a href="http://tools.ietf.org/html/rfc2046#section-5.1">RFC 2046
 333  * s5.1</a>.  @p callback is called for each part (not yet decoded in any way)
 334  * in succession; you should probably call mime_parse() for each part.
 335  */
 336 int mime_multipart(const char *s,
 337                    int (*callback)(const char *s, void *u),
 338                    const char *boundary,
 339                    void *u) {
 340   size_t bl = strlen(boundary);
 341   const char *start, *e;
 342   int ret;
 343
 344   /* We must start with a boundary string */
 345   if(!isboundary(s, boundary, bl))
 346     return -1;
 347   /* Keep going until we hit a final boundary */
 348   while(!isfinal(s, boundary, bl)) {
 349     s = strstr(s, "\r\n") + 2;
 350     start = s;
 351     while(!isboundary(s, boundary, bl)) {
 352       if(!(e = strstr(s, "\r\n")))
 353         return -1;
 354       s = e + 2;
 355     }
 356     if((ret = callback(xstrndup(start,
 357                                 s == start ? 0 : s - start - 2),
 358                        u)))
 359       return ret;
 360   }
 361   return 0;
 362 }
 363
 364 /** @brief Parse an RFC2388-style content-disposition field
 365  * @param s Start of field
 366  * @param dispositionp Where to store disposition
 367  * @param parameternamep Where to store parameter name
 368  * @param parametervaluep Wher to store parameter value
 369  * @return 0 on success, non-0 on error
 370  *
 371  * See <a href="http://tools.ietf.org/html/rfc2388#section-3">RFC 2388 s3</a>
 372  * and <a href="http://tools.ietf.org/html/rfc2183">RFC 2183</a>.
 373  */
 374 int mime_rfc2388_content_disposition(const char *s,
 375                                      char **dispositionp,
 376                                      char **parameternamep,
 377                                      char **parametervaluep) {
 378   struct dynstr disposition, parametername;
 379
 380   dynstr_init(&disposition);
 381   if(!(s = skipwhite(s, 1))) return -1;
 382   if(!*s) return -1;
 383   while(*s && !tspecial(*s) && !whitespace(*s))
 384     dynstr_append(&disposition, tolower((unsigned char)*s++));
 385   if(!(s = skipwhite(s, 1))) return -1;
 386
 387   if(*s == ';') {
 388     dynstr_init(&parametername);
 389     ++s;
 390     if(!(s = skipwhite(s, 1))) return -1;
 391     if(!*s) return -1;
 392     while(*s && !tspecial(*s) && !whitespace(*s))
 393       dynstr_append(&parametername, tolower((unsigned char)*s++));
 394     if(!(s = skipwhite(s, 1))) return -1;
 395     if(*s++ != '=') return -1;
 396     if(!(s = skipwhite(s, 1))) return -1;
 397     if(!(s = parseword(s, parametervaluep, tspecial))) return -1;
 398     if(!(s = skipwhite(s, 1))) return -1;
 399     dynstr_terminate(&parametername);
 400     *parameternamep = parametername.vec;
 401   } else
 402     *parametervaluep = *parameternamep = 0;
 403   dynstr_terminate(&disposition);
 404   *dispositionp = disposition.vec;
 405   return 0;
 406 }
 407
 408 /** @brief Convert MIME quoted-printable
 409  * @param s Quoted-printable data
 410  * @return Decoded data
 411  *
 412  * See <a href="http://tools.ietf.org/html/rfc2045#section-6.7">RFC 2045
 413  * s6.7</a>.
 414  */
 415 char *mime_qp(const char *s) {
 416   struct dynstr d;
 417   int c, a, b;
 418   const char *t;
 419
 420   dynstr_init(&d);
 421   while((c = *s++)) {
 422     switch(c) {
 423     case '=':
 424       if((a = unhexdigitq(s[0])) != -1
 425          && (b = unhexdigitq(s[1])) != -1) {
 426         dynstr_append(&d, a * 16 + b);
 427         s += 2;
 428       } else {
 429         t = s;
 430         while(*t == ' ' || *t == '\t') ++t;
 431         if(iscrlf(t)) {
 432           /* soft line break */
 433           s = t + 2;
 434         } else
 435           return 0;
 436       }
 437       break;
 438     case ' ':
 439     case '\t':
 440       t = s;
 441       while(*t == ' ' || *t == '\t') ++t;
 442       if(iscrlf(t))
 443         /* trailing space is always eliminated */
 444         s = t;
 445       else
 446         dynstr_append(&d, c);
 447       break;
 448     default:
 449       dynstr_append(&d, c);
 450       break;
 451     }
 452   }
 453   dynstr_terminate(&d);
 454   return d.vec;
 455 }
 456
 457 /** @brief Parse a RFC2109 Cookie: header
 458  * @param s Header field value
 459  * @param cd Where to store result
 460  * @return 0 on success, non-0 on error
 461  *
 462  * See <a href="http://tools.ietf.org/html/rfc2109">RFC 2109</a>.
 463  */
 464 int parse_cookie(const char *s,
 465                  struct cookiedata *cd) {
 466   char *n = 0, *v = 0;
 467
 468   memset(cd, 0, sizeof *cd);
 469   s = skipwhite(s, 0);
 470   while(*s) {
 471     /* Skip separators */
 472     if(*s == ';' || *s == ',') {
 473       ++s;
 474       s = skipwhite(s, 0);
 475       continue;
 476     }
 477     if(!(s = parsetoken(s, &n, http_separator))) return -1;
 478     s = skipwhite(s, 0);
 479     if(*s++ != '=') return -1;
 480     s = skipwhite(s, 0);
 481     if(!(s = parseword(s, &v, http_separator))) return -1;
 482     if(n[0] == '$') {
 483       /* Some bit of meta-information */
 484       if(!strcmp(n, "$Version"))
 485         cd->version = v;
 486       else if(!strcmp(n, "$Path")) {
 487         if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].path == 0)
 488           cd->cookies[cd->ncookies-1].path = v;
 489         else {
 490           error(0, "redundant $Path in Cookie: header");
 491           return -1;
 492         }
 493       } else if(!strcmp(n, "$Domain")) {
 494         if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].domain == 0)
 495           cd->cookies[cd->ncookies-1].domain = v;
 496         else {
 497           error(0, "redundant $Domain in Cookie: header");
 498           return -1;
 499         }
 500       }
 501     } else {
 502       /* It's a new cookie */
 503       cd->cookies = xrealloc(cd->cookies,
 504                              (cd->ncookies + 1) * sizeof (struct cookie));
 505       cd->cookies[cd->ncookies].name = n;
 506       cd->cookies[cd->ncookies].value = v;
 507       cd->cookies[cd->ncookies].path = 0;
 508       cd->cookies[cd->ncookies].domain = 0;
 509       ++cd->ncookies;
 510     }
 511     s = skipwhite(s, 0);
 512     if(*s && (*s != ',' && *s != ';')) {
 513       error(0, "missing separator in Cookie: header");
 514       return -1;
 515     }
 516   }
 517   return 0;
 518 }
 519
 520 /** @brief Find a named cookie
 521  * @param cd Parse cookie data
 522  * @param name Name of cookie
 523  * @return Cookie structure or NULL if not found
 524  */
 525 const struct cookie *find_cookie(const struct cookiedata *cd,
 526                                  const char *name) {
 527   int n;
 528
 529   for(n = 0; n < cd->ncookies; ++n)
 530     if(!strcmp(cd->cookies[n].name, name))
 531       return &cd->cookies[n];
 532   return 0;
 533 }
 534
 535 /** @brief RFC822 quoting
 536  * @param s String to quote
 537  * @param force If non-0, always quote
 538  * @return Possibly quoted string
 539  */
 540 char *quote822(const char *s, int force) {
 541   const char *t;
 542   struct dynstr d[1];
 543   int c;
 544
 545   if(!force) {
 546     /* See if we need to quote */
 547     for(t = s; (c = (unsigned char)*t); ++t) {
 548       if(tspecial(c) || http_separator(c) || whitespace(c))
 549         break;
 550     }
 551     if(*t)
 552       force = 1;
 553   }
 554
 555   if(!force)
 556     return xstrdup(s);
 557
 558   dynstr_init(d);
 559   dynstr_append(d, '"');
 560   for(t = s; (c = (unsigned char)*t); ++t) {
 561     if(c == '"' || c == '\\')
 562       dynstr_append(d, '\\');
 563     dynstr_append(d, c);
 564   }
 565   dynstr_append(d, '"');
 566   dynstr_terminate(d);
 567   return d->vec;
 568 }
 569
 570 /** @brief Return true if @p ptr points at trailing space */
 571 static int is_trailing_space(const char *ptr) {
 572   if(*ptr == ' ' || *ptr == '\t') {
 573     while(*ptr == ' ' || *ptr == '\t')
 574       ++ptr;
 575     return *ptr == '\n' || *ptr == 0;
 576   } else
 577     return 0;
 578 }
 579
 580 /** @brief Encoding text as quoted-printable
 581  * @param text String to encode
 582  * @return Encoded string
 583  *
 584  * See <a href="http://tools.ietf.org/html/rfc2045#section-6.7">RFC2045
 585  * s6.7</a>.
 586  */
 587 char *mime_to_qp(const char *text) {
 588   struct dynstr d[1];
 589   int linelength = 0;                   /* length of current line */
 590   char buffer[10];
 591
 592   dynstr_init(d);
 593   /* The rules are:
 594    * 1. Anything except newline can be replaced with =%02X
 595    * 2. Newline, 33-60 and 62-126 stand for themselves (i.e. not '=')
 596    * 3. Non-trailing space/tab stand for themselves.
 597    * 4. Output lines are limited to 76 chars, with =<newline> being used
 598    *    as a soft line break
 599    * 5. Newlines aren't counted towards the 76 char limit.
 600    */
 601   while(*text) {
 602     const int c = (unsigned char)*text;
 603     if(c == '\n') {
 604       /* Newline stands as itself */
 605       dynstr_append(d, '\n');
 606       linelength = 0;
 607     } else if((c >= 33 && c <= 126 && c != '=')
 608               || ((c == ' ' || c == '\t')
 609                   && !is_trailing_space(text))) {
 610       /* Things that can stand for themselves  */
 611       dynstr_append(d, c);
 612       ++linelength;
 613     } else {
 614       /* Anything else that needs encoding */
 615       snprintf(buffer, sizeof buffer, "=%02X", c);
 616       dynstr_append_string(d, buffer);
 617       linelength += 3;
 618     }
 619     ++text;
 620     if(linelength > 73 && *text && *text != '\n') {
 621       /* Next character might overflow 76 character limit if encoded, so we
 622        * insert a soft break */
 623       dynstr_append_string(d, "=\n");
 624       linelength = 0;
 625     }
 626   }
 627   /* Ensure there is a final newline */
 628   if(linelength)
 629     dynstr_append(d, '\n');
 630   /* That's all */
 631   dynstr_terminate(d);
 632   return d->vec;
 633 }
 634
 635 /** @brief Encode text
 636  * @param text Underlying UTF-8 text
 637  * @param charsetp Where to store charset string
 638  * @param encodingp Where to store encoding string
 639  * @return Encoded text (might be @ref text)
 640  */
 641 const char *mime_encode_text(const char *text,
 642                              const char **charsetp,
 643                              const char **encodingp) {
 644   const char *ptr;
 645
 646   /* See if there are in fact any non-ASCII characters */
 647   for(ptr = text; *ptr; ++ptr)
 648     if((unsigned char)*ptr >= 128)
 649       break;
 650   if(!*ptr) {
 651     /* Plain old ASCII, no encoding required */
 652     *charsetp = "us-ascii";
 653     *encodingp = "7bit";
 654     return text;
 655   }
 656   *charsetp = "utf-8";
 657   *encodingp = "quoted-printable";
 658   return mime_to_qp(text);
 659 }
 660
 661 /*
 662 Local Variables:
 663 c-basic-offset:2
 664 comment-column:40
 665 fill-column:79
 666 End:
 667 */