webcit/html2html.c

   1 /*
   2  * $Id$
   3  */
   4 /**
   5  * \defgroup HTML2HTML Output an HTML message, modifying it slightly to make sure it plays nice
   6  * with the rest of our web framework.
   7  * \ingroup WebcitHttpServer
   8  */
   9 /*@{*/
  10 #include "webcit.h"
  11 #include "webserver.h"
  12
  13
  14 /**
  15  * \brief       Strip surrounding single or double quotes from a string.
  16  *
  17  * \param s     String to be stripped.
  18  */
  19 void stripquotes(char *s)
  20 {
  21         int len;
  22
  23         if (!s) return;
  24
  25         len = strlen(s);
  26         if (len < 2) return;
  27
  28         if ( ( (s[0] == '\"') && (s[len-1] == '\"') ) || ( (s[0] == '\'') && (s[len-1] == '\'') ) ) {
  29                 s[len-1] = 0;
  30                 strcpy(s, &s[1]);
  31         }
  32 }
  33
  34
  35 /**
  36  * \brief Check to see if a META tag has overridden the declared MIME character set.
  37  *
  38  * \param charset               Character set name (left unchanged if we don't do anything)
  39  * \param meta_http_equiv       Content of the "http-equiv" portion of the META tag
  40  * \param meta_content          Content of the "content" portion of the META tag
  41  */
  42 void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_content)
  43 {
  44         char *ptr;
  45         char buf[64];
  46
  47         if (!charset) return;
  48         if (!meta_http_equiv) return;
  49         if (!meta_content) return;
  50
  51
  52         if (strcasecmp(meta_http_equiv, "Content-type")) return;
  53
  54         ptr = strchr(meta_content, ';');
  55         if (!ptr) return;
  56
  57         safestrncpy(buf, ++ptr, sizeof buf);
  58         striplt(buf);
  59         if (!strncasecmp(buf, "charset=", 8)) {
  60                 strcpy(charset, &buf[8]);
  61
  62                 /*
  63                  * The brain-damaged webmail program in Microsoft Exchange declares
  64                  * a charset of "unicode" when they really mean "UTF-8".  GNU iconv
  65                  * treats "unicode" as an alias for "UTF-16" so we have to manually
  66                  * fix this here, otherwise messages generated in Exchange webmail
  67                  * show up as a big pile of weird characters.
  68                  */
  69                 if (!strcasecmp(charset, "unicode")) {
  70                         strcpy(charset, "UTF-8");
  71                 }
  72
  73                 /* Remove wandering punctuation */
  74                 if ((ptr=strchr(charset, '\"'))) *ptr = 0;
  75                 striplt(charset);
  76         }
  77 }
  78
  79
  80
  81 /**
  82  * \brief Sanitize and enhance an HTML message for display.
  83  *        Also convert weird character sets to UTF-8 if necessary.
  84  *        Also fixup img src="cid:..." type inline images to fetch the image
  85  *
  86  * \param supplied_charset the input charset as declared in the MIME headers
  87  */
  88 void output_html(const char *supplied_charset, int treat_as_wiki, int msgnum, StrBuf *Source, StrBuf *Target) {
  89         char buf[SIZ];
  90         char *msg;
  91         char *ptr;
  92         char *msgstart;
  93         char *msgend;
  94         StrBuf *converted_msg;
  95         int buffer_length = 1;
  96         int line_length = 0;
  97         int content_length = 0;
  98         char new_window[SIZ];
  99         int brak = 0;
 100         int alevel = 0;
 101         int scriptlevel = 0;
 102         int script_start_pos = (-1);
 103         int i;
 104         int linklen;
 105         char charset[128];
 106         StrBuf *BodyArea = NULL;
 107 #ifdef HAVE_ICONV
 108         iconv_t ic = (iconv_t)(-1) ;
 109         char *ibuf;                   /**< Buffer of characters to be converted */
 110         char *obuf;                   /**< Buffer for converted characters      */
 111         size_t ibuflen;               /**< Length of input buffer               */
 112         size_t obuflen;               /**< Length of output buffer              */
 113         char *osav;                   /**< Saved pointer to output buffer       */
 114 #endif
 115         if (Target == NULL)
 116                 Target = WC->WBuf;
 117
 118         safestrncpy(charset, supplied_charset, sizeof charset);
 119         msg = strdup("");
 120         sprintf(new_window, "<a target=\"%s\" href=", TARGET);
 121
 122         if (Source == NULL) while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
 123                 line_length = strlen(buf);
 124                 buffer_length = content_length + line_length + 2;
 125                 ptr = realloc(msg, buffer_length);
 126                 if (ptr == NULL) {
 127                         StrBufAppendPrintf(Target, "<b>");
 128                         StrBufAppendPrintf(Target, _("realloc() error! couldn't get %d bytes: %s"),
 129                                         buffer_length + 1,
 130                                         strerror(errno));
 131                         StrBufAppendPrintf(Target, "</b><br /><br />\n");
 132                         while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
 133                                 /** flush */
 134                         }
 135                         free(msg);
 136                         return;
 137                 }
 138                 msg = ptr;
 139                 strcpy(&msg[content_length], buf);
 140                 content_length += line_length;
 141                 strcpy(&msg[content_length], "\n");
 142                 content_length += 1;
 143         }
 144         else {
 145                 content_length = StrLength(Source);
 146                 free(msg);
 147                 msg = (char*) ChrPtr(Source);/* TODO: remove cast */
 148                 buffer_length = content_length;
 149         }
 150
 151         /** Do a first pass to isolate the message body */
 152         ptr = msg + 1;
 153         msgstart = msg;
 154         msgend = &msg[content_length];
 155
 156         while (ptr < msgend) {
 157
 158                 /** Advance to next tag */
 159                 ptr = strchr(ptr, '<');
 160                 if ((ptr == NULL) || (ptr >= msgend)) break;
 161                 ++ptr;
 162                 if ((ptr == NULL) || (ptr >= msgend)) break;
 163
 164                 /**
 165                  *  Look for META tags.  Some messages (particularly in
 166                  *  Asian locales) illegally declare a message's character
 167                  *  set in the HTML instead of in the MIME headers.  This
 168                  *  is wrong but we have to work around it anyway.
 169                  */
 170                 if (!strncasecmp(ptr, "META", 4)) {
 171
 172                         char *meta_start;
 173                         char *meta_end;
 174                         int meta_length;
 175                         char *meta;
 176                         char *meta_http_equiv;
 177                         char *meta_content;
 178                         char *spaceptr;
 179
 180                         meta_start = &ptr[4];
 181                         meta_end = strchr(ptr, '>');
 182                         if ((meta_end != NULL) && (meta_end <= msgend)) {
 183                                 meta_length = meta_end - meta_start + 1;
 184                                 meta = malloc(meta_length + 1);
 185                                 safestrncpy(meta, meta_start, meta_length);
 186                                 meta[meta_length] = 0;
 187                                 striplt(meta);
 188                                 if (!strncasecmp(meta, "HTTP-EQUIV=", 11)) {
 189                                         meta_http_equiv = strdup(&meta[11]);
 190                                         spaceptr = strchr(meta_http_equiv, ' ');
 191                                         if (spaceptr != NULL) {
 192                                                 *spaceptr = 0;
 193                                                 meta_content = strdup(++spaceptr);
 194                                                 if (!strncasecmp(meta_content, "content=", 8)) {
 195                                                         strcpy(meta_content, &meta_content[8]);
 196                                                         stripquotes(meta_http_equiv);
 197                                                         stripquotes(meta_content);
 198                                                         extract_charset_from_meta(charset,
 199                                                                         meta_http_equiv, meta_content);
 200                                                 }
 201                                                 free(meta_content);
 202                                         }
 203                                         free(meta_http_equiv);
 204                                 }
 205                                 free(meta);
 206                         }
 207                 }
 208
 209                 /**
 210                  * Any of these tags cause everything up to and including
 211                  * the tag to be removed.
 212                  */
 213                 if ( (!strncasecmp(ptr, "HTML", 4))
 214                                 ||(!strncasecmp(ptr, "HEAD", 4))
 215                                 ||(!strncasecmp(ptr, "/HEAD", 5))
 216                                 ||(!strncasecmp(ptr, "BODY", 4)) ) {
 217                         char *pBody = NULL;
 218
 219                         if (!strncasecmp(ptr, "BODY", 4)) {
 220                                 pBody = ptr;
 221                         }
 222                         ptr = strchr(ptr, '>');
 223                         if ((ptr == NULL) || (ptr >= msgend)) break;
 224                         if ((pBody != NULL) && (ptr - pBody > 4)) {
 225                                 char* src;
 226                                 char *cid_start, *cid_end;
 227
 228                                 *ptr = '\0';
 229                                 pBody += 4;
 230                                 while ((isspace(*pBody)) && (pBody < ptr))
 231                                         pBody ++;
 232                                 BodyArea = NewStrBufPlain(NULL,  ptr - pBody);
 233
 234                                 if (pBody < ptr) {
 235                                         src = strstr(pBody, "cid:");
 236                                         if (src) {
 237                                                 cid_start = src + 4;
 238                                                 cid_end = cid_start;
 239                                                 while ((*cid_end != '"') &&
 240                                                                 !isspace(*cid_end) &&
 241                                                                 (cid_end < ptr))
 242                                                         cid_end ++;
 243
 244                                                 /* copy tag and attributes up to src="cid: */
 245                                                 StrBufAppendBufPlain(BodyArea, pBody, src - pBody, 0);
 246
 247                                                 /* add in /webcit/mimepart/<msgno>/CID/
 248                                                    trailing / stops dumb URL filters getting excited */
 249                                                 StrBufAppendPrintf(BodyArea,
 250                                                                 "/webcit/mimepart/%d/",msgnum);
 251                                                 StrBufAppendBufPlain(BodyArea, cid_start, cid_end - cid_start, 0);
 252
 253                                                 if (ptr - cid_end > 0)
 254                                                         StrBufAppendBufPlain(BodyArea,
 255                                                                         cid_end + 1,
 256                                                                         ptr - cid_end, 0);
 257                                         }
 258                                         else
 259                                                 StrBufAppendBufPlain(BodyArea, pBody, ptr - pBody, 0);
 260                                 }
 261                                 *ptr = '>';
 262                         }
 263                         ++ptr;
 264                         if ((ptr == NULL) || (ptr >= msgend)) break;
 265                         msgstart = ptr;
 266                 }
 267
 268                 /**
 269                  * Any of these tags cause everything including and following
 270                  * the tag to be removed.
 271                  */
 272                 if ( (!strncasecmp(ptr, "/HTML", 5))
 273                                 ||(!strncasecmp(ptr, "/BODY", 5)) ) {
 274                         --ptr;
 275                         msgend = ptr;
 276                         strcpy(ptr, "");
 277
 278                 }
 279
 280                 ++ptr;
 281         }
 282         if (msgstart > msg) {
 283                 strcpy(msg, msgstart);
 284         }
 285
 286         /** Now go through the message, parsing tags as necessary. */
 287         converted_msg = NewStrBufPlain(NULL, content_length + 8192);
 288
 289
 290         /** Convert foreign character sets to UTF-8 if necessary. */
 291 #ifdef HAVE_ICONV
 292         if ( (strcasecmp(charset, "us-ascii"))
 293                         && (strcasecmp(charset, "UTF-8"))
 294                         && (strcasecmp(charset, ""))
 295            ) {
 296                 lprintf(9, "Converting %s to UTF-8\n", charset);
 297                 ctdl_iconv_open("UTF-8", charset, &ic);
 298                 if (ic == (iconv_t)(-1) ) {
 299                         lprintf(5, "%s:%d iconv_open() failed: %s\n",
 300                                         __FILE__, __LINE__, strerror(errno));
 301                 }
 302         }
 303         if  (Source == NULL) {
 304                 if (ic != (iconv_t)(-1) ) {
 305                         ibuf = msg;
 306                         ibuflen = content_length;
 307                         obuflen = content_length + (content_length / 2) ;
 308                         obuf = (char *) malloc(obuflen);
 309                         osav = obuf;
 310                         iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
 311                         content_length = content_length + (content_length / 2) - obuflen;
 312                         osav[content_length] = 0;
 313                         free(msg);
 314                         msg = osav;
 315                         iconv_close(ic);
 316                 }
 317         }
 318         else {
 319                 if (ic != (iconv_t)(-1) ) {
 320                         StrBuf *Buf = NewStrBufPlain(NULL, StrLength(Source) + 8096);;
 321                         StrBufConvert(Source, Buf, &ic);
 322                         FreeStrBuf(&Buf);
 323                         iconv_close(ic);
 324                         msg = (char*)ChrPtr(Source); /* TODO: get rid of this. */
 325                 }
 326         }
 327
 328 #endif
 329
 330         /**
 331          *      At this point, the message has been stripped down to
 332          *      only the content inside the <BODY></BODY> tags, and has
 333          *      been converted to UTF-8 if it was originally in a foreign
 334          *      character set.  The text is also guaranteed to be null
 335          *      terminated now.
 336          */
 337
 338         if (converted_msg == NULL) {
 339                 StrBufAppendPrintf(Target, "Error %d: %s<br />%s:%d", errno, strerror(errno), __FILE__, __LINE__);
 340                 goto BAIL;
 341         }
 342
 343         if (BodyArea != NULL) {
 344                 StrBufAppendBufPlain(converted_msg, HKEY("<table "), 0);
 345                 StrBufAppendBuf(converted_msg, BodyArea, 0);
 346                 StrBufAppendBufPlain(converted_msg, HKEY(" width=\"100%\"><tr><td>"), 0);
 347         }
 348         ptr = msg;
 349         msgend = strchr(msg, 0);
 350         while (ptr < msgend) {
 351
 352                 /** Try to sanitize the html of any rogue scripts */
 353                 if (!strncasecmp(ptr, "<script", 7)) {
 354                         if (scriptlevel == 0) {
 355                                 script_start_pos = StrLength(converted_msg);
 356                         }
 357                         ++scriptlevel;
 358                 }
 359                 if (!strncasecmp(ptr, "</script", 8)) {
 360                         --scriptlevel;
 361                 }
 362
 363                 /**
 364                  * Change mailto: links to WebCit mail, by replacing the
 365                  * link with one that points back to our mail room.  Due to
 366                  * the way we parse URL's, it'll even handle mailto: links
 367                  * that have "?subject=" in them.
 368                  */
 369                 if (!strncasecmp(ptr, "<a href=\"mailto:", 16)) {
 370                         content_length += 64;
 371                         StrBufAppendPrintf(converted_msg,
 372                                         "<a href=\"display_enter?force_room=_MAIL_&recp=");
 373                         ptr = &ptr[16];
 374                         ++alevel;
 375                         ++brak;
 376                 }
 377                 /** Make external links open in a separate window */
 378                 else if (!strncasecmp(ptr, "<a href=\"", 9)) {
 379                         ++alevel;
 380                         ++brak;
 381                         if ( ((strchr(ptr, ':') < strchr(ptr, '/')))
 382                                         &&  ((strchr(ptr, '/') < strchr(ptr, '>')))
 383                            ) {
 384                                 /* open external links to new window */
 385                                 StrBufAppendPrintf(converted_msg, new_window);
 386                                 ptr = &ptr[8];
 387                         }
 388                         else if ( (treat_as_wiki) && (strncasecmp(ptr, "<a href=\"wiki?", 14)) ) {
 389                                 content_length += 64;
 390                                 StrBufAppendPrintf(converted_msg, "<a href=\"wiki?page=");
 391                                 ptr = &ptr[9];
 392                         }
 393                         else {
 394                                 StrBufAppendPrintf(converted_msg, "<a href=\"");
 395                                 ptr = &ptr[9];
 396                         }
 397                 }
 398                 /** Fixup <img src="cid:... ...> to fetch the mime part */
 399                 else if (!strncasecmp(ptr, "<img ", 5)) {
 400                         char* tag_end=strchr(ptr,'>');
 401
 402                         /* FIXME - handle this situation (maybe someone opened an <img cid...
 403                          * and then ended the message)
 404                          */
 405                         if (!tag_end) {
 406                                 lprintf(9, "tag_end is null and ptr is:\n");
 407                                 lprintf(9, "%s\n", ptr);
 408                                 lprintf(9, "Theoretical bytes remaining: %d\n", msgend - ptr);
 409                         }
 410
 411                         char* src=strstr(ptr, " src=\"cid:");
 412                         char *cid_start, *cid_end;
 413                         ++brak;
 414
 415                         if (src
 416                                 && tag_end
 417                                 && (cid_start=strchr(src,':'))
 418                                 && (cid_end=strchr(cid_start,'"'))
 419                                 && (cid_end < tag_end)
 420                         ) {
 421                                 /* copy tag and attributes up to src="cid: */
 422                                 StrBufAppendBufPlain(converted_msg, ptr, src - ptr, 0);
 423                                 cid_start++;
 424
 425                                 /* add in /webcit/mimepart/<msgno>/CID/
 426                                    trailing / stops dumb URL filters getting excited */
 427                                 StrBufAppendPrintf(converted_msg,
 428                                                 " src=\"/webcit/mimepart/%d/",msgnum);
 429                                 StrBufAppendBufPlain(converted_msg, cid_start, cid_end - cid_start, 0);
 430                                 StrBufAppendBufPlain(converted_msg, "/\"", -1, 0);
 431
 432                                 ptr = cid_end+1;
 433                         }
 434                         StrBufAppendBufPlain(converted_msg, ptr, tag_end - ptr, 0);
 435                         ptr = tag_end;
 436                 }
 437
 438                 /**
 439                  * Turn anything that looks like a URL into a real link, as long
 440                  * as it's not inside a tag already
 441                  */
 442                 else if ( (brak == 0) && (alevel == 0)
 443                      && (!strncasecmp(ptr, "http://", 7))) {
 444                                 /** Find the end of the link */
 445                                 int strlenptr;
 446                                 linklen = 0;
 447
 448                                 strlenptr = strlen(ptr);
 449                                 for (i=0; i<=strlenptr; ++i) {
 450                                         if ((ptr[i]==0)
 451                                            ||(isspace(ptr[i]))
 452                                            ||(ptr[i]==10)
 453                                            ||(ptr[i]==13)
 454                                            ||(ptr[i]=='(')
 455                                            ||(ptr[i]==')')
 456                                            ||(ptr[i]=='<')
 457                                            ||(ptr[i]=='>')
 458                                            ||(ptr[i]=='[')
 459                                            ||(ptr[i]==']')
 460                                            ||(ptr[i]=='"')
 461                                            ||(ptr[i]=='\'')
 462                                         ) linklen = i;
 463                                         /* did s.b. send us an entity? */
 464                                         if (ptr[i] == '&') {
 465                                                 if ((ptr[i+2] ==';') ||
 466                                                     (ptr[i+3] ==';') ||
 467                                                     (ptr[i+5] ==';') ||
 468                                                     (ptr[i+6] ==';') ||
 469                                                     (ptr[i+7] ==';'))
 470                                                         linklen = i;
 471                                         }
 472                                         if (linklen > 0) break;
 473                                 }
 474                                 if (linklen > 0) {
 475                                         char *ltreviewptr;
 476                                         char *nbspreviewptr;
 477                                         char linkedchar;
 478                                         int len;
 479
 480                                         len = linklen;
 481                                         linkedchar = ptr[len];
 482                                         ptr[len] = '\0';
 483                                         /* spot for some subject strings tinymce tends to give us. */
 484                                         ltreviewptr = strchr(ptr, '<');
 485                                         if (ltreviewptr != NULL) {
 486                                                 *ltreviewptr = '\0';
 487                                                 linklen = ltreviewptr - ptr;
 488                                         }
 489
 490                                         nbspreviewptr = strstr(ptr, "&nbsp;");
 491                                         if (nbspreviewptr != NULL) {
 492                                                 /* nbspreviewptr = '\0'; */
 493                                                 linklen = nbspreviewptr - ptr;
 494                                         }
 495                                         if (ltreviewptr != 0)
 496                                                 *ltreviewptr = '<';
 497
 498                                         ptr[len] = linkedchar;
 499
 500                                         content_length += (32 + linklen);
 501                                         StrBufAppendPrintf(converted_msg, "%s\"", new_window);
 502                                         StrBufAppendBufPlain(converted_msg, ptr, linklen, 0);
 503                                         StrBufAppendPrintf(converted_msg, "\">");
 504                                         StrBufAppendBufPlain(converted_msg, ptr, linklen, 0);
 505                                         ptr += linklen;
 506                                         StrBufAppendPrintf(converted_msg, "</A>");
 507                                 }
 508                 }
 509                 else {
 510                         StrBufAppendBufPlain(converted_msg, ptr, 1, 0);
 511                         ptr++;
 512                 }
 513
 514
 515                 if ((ptr >= msg) && (ptr <= msgend)) {
 516                         /*
 517                          * We need to know when we're inside a tag,
 518                          * so we don't turn things that look like URL's into
 519                          * links, when they're already links - or image sources.
 520                          */
 521                         if ((ptr > msg) && (*(ptr-1) == '<')) {
 522                                 ++brak;
 523                         }
 524                         if ((ptr > msg) && (*(ptr-1) == '>')) {
 525                                 --brak;
 526                                 if ((scriptlevel == 0) && (script_start_pos >= 0)) {
 527                                         StrBufCutRight(converted_msg, StrLength(converted_msg) - script_start_pos);
 528                                         script_start_pos = (-1);
 529                                 }
 530                         }
 531                         if (!strncasecmp(ptr, "</A>", 3)) --alevel;
 532                 }
 533         }
 534
 535         if (BodyArea != NULL) {
 536                 StrBufAppendBufPlain(converted_msg, HKEY("</td></tr></table>"), 0);
 537                 FreeStrBuf(&BodyArea);
 538         }
 539
 540         /**     uncomment these two lines to override conversion        */
 541         /**     memcpy(converted_msg, msg, content_length);             */
 542         /**     output_length = content_length;                         */
 543
 544         /** Output our big pile of markup */
 545         StrBufAppendBuf(Target, converted_msg, 0);
 546
 547 BAIL:   /** A little trailing vertical whitespace... */
 548         StrBufAppendPrintf(Target, "<br /><br />\n");
 549
 550         /** Now give back the memory */
 551         FreeStrBuf(&converted_msg);
 552         if ((msg != NULL) && (Source == NULL)) free(msg);
 553 }
 554
 555
 556
 557
 558
 559
 560 /*
 561  * Look for URL's embedded in a buffer and make them linkable.  We use a
 562  * target window in order to keep the Citadel session in its own window.
 563  */
 564 void UrlizeText(StrBuf* Target, StrBuf *Source, StrBuf *WrkBuf)
 565 {
 566         int len, UrlLen, Offset, TrailerLen;
 567         const char *start, *end, *pos;
 568
 569         FlushStrBuf(Target);
 570
 571         start = NULL;
 572         len = StrLength(Source);
 573         end = ChrPtr(Source) + len;
 574         for (pos = ChrPtr(Source); (pos < end) && (start == NULL); ++pos) {
 575                 if (!strncasecmp(pos, "http://", 7))
 576                         start = pos;
 577                 else if (!strncasecmp(pos, "ftp://", 6))
 578                         start = pos;
 579         }
 580
 581         if (start == NULL) {
 582                 StrBufAppendBuf(Target, Source, 0);
 583                 return;
 584         }
 585         FlushStrBuf(WrkBuf);
 586
 587         for (pos = ChrPtr(Source) + len; pos > start; --pos) {
 588                 if (  (!isprint(*pos))
 589                    || (isspace(*pos))
 590                    || (*pos == '{')
 591                    || (*pos == '}')
 592                    || (*pos == '|')
 593                    || (*pos == '\\')
 594                    || (*pos == '^')
 595                    || (*pos == '[')
 596                    || (*pos == ']')
 597                    || (*pos == '`')
 598                    || (*pos == '<')
 599                    || (*pos == '>')
 600                    || (*pos == '(')
 601                    || (*pos == ')')
 602                 ) {
 603                         end = pos;
 604                 }
 605         }
 606
 607         UrlLen = end - start;
 608         StrBufAppendBufPlain(WrkBuf, start, UrlLen, 0);
 609
 610         Offset = start - ChrPtr(Source);
 611         if (Offset != 0)
 612                 StrBufAppendBufPlain(Target, ChrPtr(Source), Offset, 0);
 613         StrBufAppendPrintf(Target, "%ca href=%c%s%c TARGET=%c%s%c%c%s%c/A%c",
 614                            LB, QU, ChrPtr(WrkBuf), QU, QU, TARGET,
 615                            QU, RB, ChrPtr(WrkBuf), LB, RB);
 616
 617         TrailerLen = StrLength(Source) - (end - ChrPtr(Source));
 618         if (TrailerLen > 0)
 619                 StrBufAppendBufPlain(Target, end, TrailerLen, 0);
 620 }
 621 void url(char *buf, size_t bufsize)
 622 {
 623         int len, UrlLen, Offset, TrailerLen, outpos;
 624         char *start, *end, *pos;
 625         char urlbuf[SIZ];
 626         char outbuf[SIZ];
 627
 628         start = NULL;
 629         len = strlen(buf);
 630         if (len > bufsize) {
 631                 lprintf(1, "URL: content longer than buffer!");
 632                 return;
 633         }
 634         end = buf + len;
 635         for (pos = buf; (pos < end) && (start == NULL); ++pos) {
 636                 if (!strncasecmp(pos, "http://", 7))
 637                         start = pos;
 638                 if (!strncasecmp(pos, "ftp://", 6))
 639                         start = pos;
 640         }
 641
 642         if (start == NULL)
 643                 return;
 644
 645         for (pos = buf+len; pos > start; --pos) {
 646                 if (  (!isprint(*pos))
 647                    || (isspace(*pos))
 648                    || (*pos == '{')
 649                    || (*pos == '}')
 650                    || (*pos == '|')
 651                    || (*pos == '\\')
 652                    || (*pos == '^')
 653                    || (*pos == '[')
 654                    || (*pos == ']')
 655                    || (*pos == '`')
 656                    || (*pos == '<')
 657                    || (*pos == '>')
 658                    || (*pos == '(')
 659                    || (*pos == ')')
 660                 ) {
 661                         end = pos;
 662                 }
 663         }
 664
 665         UrlLen = end - start;
 666         if (UrlLen > sizeof(urlbuf)){
 667                 lprintf(1, "URL: content longer than buffer!");
 668                 return;
 669         }
 670         memcpy(urlbuf, start, UrlLen);
 671         urlbuf[UrlLen] = '\0';
 672
 673         Offset = start - buf;
 674         if ((Offset != 0) && (Offset < sizeof(outbuf)))
 675                 memcpy(outbuf, buf, Offset);
 676         outpos = snprintf(&outbuf[Offset], sizeof(outbuf) - Offset,
 677                           "%ca href=%c%s%c TARGET=%c%s%c%c%s%c/A%c",
 678                           LB, QU, urlbuf, QU, QU, TARGET, QU, RB, urlbuf, LB, RB);
 679         if (outpos >= sizeof(outbuf) - Offset) {
 680                 lprintf(1, "URL: content longer than buffer!");
 681                 return;
 682         }
 683
 684         TrailerLen = len - (end - start);
 685         if (TrailerLen > 0)
 686                 memcpy(outbuf + Offset + outpos, end, TrailerLen);
 687         if (Offset + outpos + TrailerLen > bufsize) {
 688                 lprintf(1, "URL: content longer than buffer!");
 689                 return;
 690         }
 691         memcpy (buf, outbuf, Offset + outpos + TrailerLen);
 692         *(buf + Offset + outpos + TrailerLen) = '\0';
 693 }
 694
 695
 696
 697
 698 /*@}*/