webcit/html2html.c

   1 /*
   2  * Output an HTML message, modifying it slightly to make sure it plays nice
   3  * with the rest of our web framework.
   4  *
   5  * Copyright (c) 2005-2010 by the citadel.org team
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 3 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License
  18  * along with this program; if not, write to the Free Software
  19  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  20  */
  21
  22 #include "webcit.h"
  23 #include "webserver.h"
  24
  25
  26 /*
  27  * Strip surrounding single or double quotes from a string.
  28  */
  29 void stripquotes(char *s)
  30 {
  31         int len;
  32
  33         if (!s) return;
  34
  35         len = strlen(s);
  36         if (len < 2) return;
  37
  38         if ( ( (s[0] == '\"') && (s[len-1] == '\"') ) || ( (s[0] == '\'') && (s[len-1] == '\'') ) ) {
  39                 s[len-1] = 0;
  40                 strcpy(s, &s[1]);
  41         }
  42 }
  43
  44
  45 /*
  46  * Check to see if a META tag has overridden the declared MIME character set.
  47  *
  48  * charset              Character set name (left unchanged if we don't do anything)
  49  * meta_http_equiv      Content of the "http-equiv" portion of the META tag
  50  * meta_content         Content of the "content" portion of the META tag
  51  */
  52 void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_content)
  53 {
  54         char *ptr;
  55         char buf[64];
  56
  57         if (!charset) return;
  58         if (!meta_http_equiv) return;
  59         if (!meta_content) return;
  60
  61
  62         if (strcasecmp(meta_http_equiv, "Content-type")) return;
  63
  64         ptr = strchr(meta_content, ';');
  65         if (!ptr) return;
  66
  67         safestrncpy(buf, ++ptr, sizeof buf);
  68         striplt(buf);
  69         if (!strncasecmp(buf, "charset=", 8)) {
  70                 strcpy(charset, &buf[8]);
  71
  72                 /*
  73                  * The brain-damaged webmail program in Microsoft Exchange declares
  74                  * a charset of "unicode" when they really mean "UTF-8".  GNU iconv
  75                  * treats "unicode" as an alias for "UTF-16" so we have to manually
  76                  * fix this here, otherwise messages generated in Exchange webmail
  77                  * show up as a big pile of weird characters.
  78                  */
  79                 if (!strcasecmp(charset, "unicode")) {
  80                         strcpy(charset, "UTF-8");
  81                 }
  82
  83                 /* Remove wandering punctuation */
  84                 if ((ptr=strchr(charset, '\"'))) *ptr = 0;
  85                 striplt(charset);
  86         }
  87 }
  88
  89
  90
  91 /*
  92  * Sanitize and enhance an HTML message for display.
  93  * Also convert weird character sets to UTF-8 if necessary.
  94  * Also fixup img src="cid:..." type inline images to fetch the image
  95  *
  96  */
  97 void output_html(const char *supplied_charset, int treat_as_wiki, int msgnum, StrBuf *Source, StrBuf *Target) {
  98         char buf[SIZ];
  99         char *msg;
 100         char *ptr;
 101         char *msgstart;
 102         char *msgend;
 103         StrBuf *converted_msg;
 104         int buffer_length = 1;
 105         int line_length = 0;
 106         int content_length = 0;
 107         char new_window[SIZ];
 108         int brak = 0;
 109         int alevel = 0;
 110         int scriptlevel = 0;
 111         int script_start_pos = (-1);
 112         int i;
 113         int linklen;
 114         char charset[128];
 115         StrBuf *BodyArea = NULL;
 116 #ifdef HAVE_ICONV
 117         iconv_t ic = (iconv_t)(-1) ;
 118         char *ibuf;                   /* Buffer of characters to be converted */
 119         char *obuf;                   /* Buffer for converted characters      */
 120         size_t ibuflen;               /* Length of input buffer               */
 121         size_t obuflen;               /* Length of output buffer              */
 122         char *osav;                   /* Saved pointer to output buffer       */
 123 #endif
 124         if (Target == NULL)
 125                 Target = WC->WBuf;
 126
 127         safestrncpy(charset, supplied_charset, sizeof charset);
 128         msg = strdup("");
 129         sprintf(new_window, "<a target=\"%s\" href=", TARGET);
 130
 131         if (Source == NULL) while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
 132                 line_length = strlen(buf);
 133                 buffer_length = content_length + line_length + 2;
 134                 ptr = realloc(msg, buffer_length);
 135                 if (ptr == NULL) {
 136                         StrBufAppendPrintf(Target, "<b>");
 137                         StrBufAppendPrintf(Target, _("realloc() error! couldn't get %d bytes: %s"),
 138                                         buffer_length + 1,
 139                                         strerror(errno));
 140                         StrBufAppendPrintf(Target, "</b><br /><br />\n");
 141                         while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
 142                                 /** flush */
 143                         }
 144                         free(msg);
 145                         return;
 146                 }
 147                 msg = ptr;
 148                 strcpy(&msg[content_length], buf);
 149                 content_length += line_length;
 150                 strcpy(&msg[content_length], "\n");
 151                 content_length += 1;
 152         }
 153         else {
 154                 content_length = StrLength(Source);
 155                 free(msg);
 156                 msg = (char*) ChrPtr(Source);/* TODO: remove cast */
 157                 buffer_length = content_length;
 158         }
 159
 160         /** Do a first pass to isolate the message body */
 161         ptr = msg + 1;
 162         msgstart = msg;
 163         msgend = &msg[content_length];
 164
 165         while (ptr < msgend) {
 166
 167                 /** Advance to next tag */
 168                 ptr = strchr(ptr, '<');
 169                 if ((ptr == NULL) || (ptr >= msgend)) break;
 170                 ++ptr;
 171                 if ((ptr == NULL) || (ptr >= msgend)) break;
 172
 173                 /*
 174                  *  Look for META tags.  Some messages (particularly in
 175                  *  Asian locales) illegally declare a message's character
 176                  *  set in the HTML instead of in the MIME headers.  This
 177                  *  is wrong but we have to work around it anyway.
 178                  */
 179                 if (!strncasecmp(ptr, "META", 4)) {
 180
 181                         char *meta_start;
 182                         char *meta_end;
 183                         int meta_length;
 184                         char *meta;
 185                         char *meta_http_equiv;
 186                         char *meta_content;
 187                         char *spaceptr;
 188
 189                         meta_start = &ptr[4];
 190                         meta_end = strchr(ptr, '>');
 191                         if ((meta_end != NULL) && (meta_end <= msgend)) {
 192                                 meta_length = meta_end - meta_start + 1;
 193                                 meta = malloc(meta_length + 1);
 194                                 safestrncpy(meta, meta_start, meta_length);
 195                                 meta[meta_length] = 0;
 196                                 striplt(meta);
 197                                 if (!strncasecmp(meta, "HTTP-EQUIV=", 11)) {
 198                                         meta_http_equiv = strdup(&meta[11]);
 199                                         spaceptr = strchr(meta_http_equiv, ' ');
 200                                         if (spaceptr != NULL) {
 201                                                 *spaceptr = 0;
 202                                                 meta_content = strdup(++spaceptr);
 203                                                 if (!strncasecmp(meta_content, "content=", 8)) {
 204                                                         strcpy(meta_content, &meta_content[8]);
 205                                                         stripquotes(meta_http_equiv);
 206                                                         stripquotes(meta_content);
 207                                                         extract_charset_from_meta(charset,
 208                                                                         meta_http_equiv, meta_content);
 209                                                 }
 210                                                 free(meta_content);
 211                                         }
 212                                         free(meta_http_equiv);
 213                                 }
 214                                 free(meta);
 215                         }
 216                 }
 217
 218                 /*
 219                  * Any of these tags cause everything up to and including
 220                  * the tag to be removed.
 221                  */
 222                 if ( (!strncasecmp(ptr, "HTML", 4))
 223                                 ||(!strncasecmp(ptr, "HEAD", 4))
 224                                 ||(!strncasecmp(ptr, "/HEAD", 5))
 225                                 ||(!strncasecmp(ptr, "BODY", 4)) ) {
 226                         char *pBody = NULL;
 227
 228                         if (!strncasecmp(ptr, "BODY", 4)) {
 229                                 pBody = ptr;
 230                         }
 231                         ptr = strchr(ptr, '>');
 232                         if ((ptr == NULL) || (ptr >= msgend)) break;
 233                         if ((pBody != NULL) && (ptr - pBody > 4)) {
 234                                 char* src;
 235                                 char *cid_start, *cid_end;
 236
 237                                 *ptr = '\0';
 238                                 pBody += 4;
 239                                 while ((isspace(*pBody)) && (pBody < ptr))
 240                                         pBody ++;
 241                                 BodyArea = NewStrBufPlain(NULL,  ptr - pBody);
 242
 243                                 if (pBody < ptr) {
 244                                         src = strstr(pBody, "cid:");
 245                                         if (src) {
 246                                                 cid_start = src + 4;
 247                                                 cid_end = cid_start;
 248                                                 while ((*cid_end != '"') &&
 249                                                                 !isspace(*cid_end) &&
 250                                                                 (cid_end < ptr))
 251                                                         cid_end ++;
 252
 253                                                 /* copy tag and attributes up to src="cid: */
 254                                                 StrBufAppendBufPlain(BodyArea, pBody, src - pBody, 0);
 255
 256                                                 /* add in /webcit/mimepart/<msgno>/CID/
 257                                                    trailing / stops dumb URL filters getting excited */
 258                                                 StrBufAppendPrintf(BodyArea,
 259                                                                 "/webcit/mimepart/%d/",msgnum);
 260                                                 StrBufAppendBufPlain(BodyArea, cid_start, cid_end - cid_start, 0);
 261
 262                                                 if (ptr - cid_end > 0)
 263                                                         StrBufAppendBufPlain(BodyArea,
 264                                                                         cid_end + 1,
 265                                                                         ptr - cid_end, 0);
 266                                         }
 267                                         else
 268                                                 StrBufAppendBufPlain(BodyArea, pBody, ptr - pBody, 0);
 269                                 }
 270                                 *ptr = '>';
 271                         }
 272                         ++ptr;
 273                         if ((ptr == NULL) || (ptr >= msgend)) break;
 274                         msgstart = ptr;
 275                 }
 276
 277                 /*
 278                  * Any of these tags cause everything including and following
 279                  * the tag to be removed.
 280                  */
 281                 if ( (!strncasecmp(ptr, "/HTML", 5))
 282                                 ||(!strncasecmp(ptr, "/BODY", 5)) ) {
 283                         --ptr;
 284                         msgend = ptr;
 285                         strcpy(ptr, "");
 286
 287                 }
 288
 289                 ++ptr;
 290         }
 291         if (msgstart > msg) {
 292                 strcpy(msg, msgstart);
 293         }
 294
 295         /* Now go through the message, parsing tags as necessary. */
 296         converted_msg = NewStrBufPlain(NULL, content_length + 8192);
 297
 298
 299         /** Convert foreign character sets to UTF-8 if necessary. */
 300 #ifdef HAVE_ICONV
 301         if ( (strcasecmp(charset, "us-ascii"))
 302                         && (strcasecmp(charset, "UTF-8"))
 303                         && (strcasecmp(charset, ""))
 304            ) {
 305                 lprintf(9, "Converting %s to UTF-8\n", charset);
 306                 ctdl_iconv_open("UTF-8", charset, &ic);
 307                 if (ic == (iconv_t)(-1) ) {
 308                         lprintf(5, "%s:%d iconv_open() failed: %s\n",
 309                                         __FILE__, __LINE__, strerror(errno));
 310                 }
 311         }
 312         if  (Source == NULL) {
 313                 if (ic != (iconv_t)(-1) ) {
 314                         ibuf = msg;
 315                         ibuflen = content_length;
 316                         obuflen = content_length + (content_length / 2) ;
 317                         obuf = (char *) malloc(obuflen);
 318                         osav = obuf;
 319                         iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
 320                         content_length = content_length + (content_length / 2) - obuflen;
 321                         osav[content_length] = 0;
 322                         free(msg);
 323                         msg = osav;
 324                         iconv_close(ic);
 325                 }
 326         }
 327         else {
 328                 if (ic != (iconv_t)(-1) ) {
 329                         StrBuf *Buf = NewStrBufPlain(NULL, StrLength(Source) + 8096);;
 330                         StrBufConvert(Source, Buf, &ic);
 331                         FreeStrBuf(&Buf);
 332                         iconv_close(ic);
 333                         msg = (char*)ChrPtr(Source); /* TODO: get rid of this. */
 334                 }
 335         }
 336
 337 #endif
 338
 339         /*
 340          *      At this point, the message has been stripped down to
 341          *      only the content inside the <BODY></BODY> tags, and has
 342          *      been converted to UTF-8 if it was originally in a foreign
 343          *      character set.  The text is also guaranteed to be null
 344          *      terminated now.
 345          */
 346
 347         if (converted_msg == NULL) {
 348                 StrBufAppendPrintf(Target, "Error %d: %s<br />%s:%d", errno, strerror(errno), __FILE__, __LINE__);
 349                 goto BAIL;
 350         }
 351
 352         if (BodyArea != NULL) {
 353                 StrBufAppendBufPlain(converted_msg, HKEY("<table "), 0);
 354                 StrBufAppendBuf(converted_msg, BodyArea, 0);
 355                 StrBufAppendBufPlain(converted_msg, HKEY(" width=\"100%\"><tr><td>"), 0);
 356         }
 357         ptr = msg;
 358         msgend = strchr(msg, 0);
 359         while (ptr < msgend) {
 360
 361                 /** Try to sanitize the html of any rogue scripts */
 362                 if (!strncasecmp(ptr, "<script", 7)) {
 363                         if (scriptlevel == 0) {
 364                                 script_start_pos = StrLength(converted_msg);
 365                         }
 366                         ++scriptlevel;
 367                 }
 368                 if (!strncasecmp(ptr, "</script", 8)) {
 369                         --scriptlevel;
 370                 }
 371
 372                 /**
 373                  * Change mailto: links to WebCit mail, by replacing the
 374                  * link with one that points back to our mail room.  Due to
 375                  * the way we parse URL's, it'll even handle mailto: links
 376                  * that have "?subject=" in them.
 377                  */
 378                 if (!strncasecmp(ptr, "<a href=\"mailto:", 16)) {
 379                         content_length += 64;
 380                         StrBufAppendPrintf(converted_msg,
 381                                         "<a href=\"display_enter?force_room=_MAIL_?recp=");
 382                         ptr = &ptr[16];
 383                         ++alevel;
 384                         ++brak;
 385                 }
 386                 /** Make external links open in a separate window */
 387                 else if (!strncasecmp(ptr, "<a href=\"", 9)) {
 388                         ++alevel;
 389                         ++brak;
 390                         if ( ((strchr(ptr, ':') < strchr(ptr, '/')))
 391                                         &&  ((strchr(ptr, '/') < strchr(ptr, '>')))
 392                            ) {
 393                                 /* open external links to new window */
 394                                 StrBufAppendPrintf(converted_msg, new_window);
 395                                 ptr = &ptr[8];
 396                         }
 397                         else if ( (treat_as_wiki) && (strncasecmp(ptr, "<a href=\"wiki?", 14)) ) {
 398                                 content_length += 64;
 399                                 StrBufAppendPrintf(converted_msg, "<a href=\"wiki?page=");
 400                                 ptr = &ptr[9];
 401                         }
 402                         else {
 403                                 StrBufAppendPrintf(converted_msg, "<a href=\"");
 404                                 ptr = &ptr[9];
 405                         }
 406                 }
 407                 /** Fixup <img src="cid:... ...> to fetch the mime part */
 408                 else if (!strncasecmp(ptr, "<img ", 5)) {
 409                         char *cid_start, *cid_end;
 410                         char* tag_end=strchr(ptr,'>');
 411                         char* src;
 412                         /* FIXME - handle this situation (maybe someone opened an <img cid...
 413                          * and then ended the message)
 414                          */
 415                         if (!tag_end) {
 416                                 lprintf(9, "tag_end is null and ptr is:\n");
 417                                 lprintf(9, "%s\n", ptr);
 418                                 lprintf(9, "Theoretical bytes remaining: %d\n", msgend - ptr);
 419                         }
 420
 421                         src=strstr(ptr, "src=\"cid:");
 422                         ++brak;
 423
 424                         if (src
 425                             && isspace(*(src-1))
 426                                 && tag_end
 427                                 && (cid_start=strchr(src,':'))
 428                                 && (cid_end=strchr(cid_start,'"'))
 429                                 && (cid_end < tag_end)
 430                         ) {
 431                                 /* copy tag and attributes up to src="cid: */
 432                                 StrBufAppendBufPlain(converted_msg, ptr, src - ptr, 0);
 433                                 cid_start++;
 434
 435                                 /* add in /webcit/mimepart/<msgno>/CID/
 436                                    trailing / stops dumb URL filters getting excited */
 437                                 StrBufAppendPrintf(converted_msg,
 438                                                 " src=\"/webcit/mimepart/%d/",msgnum);
 439                                 StrBufAppendBufPlain(converted_msg, cid_start, cid_end - cid_start, 0);
 440                                 StrBufAppendBufPlain(converted_msg, "/\"", -1, 0);
 441
 442                                 ptr = cid_end+1;
 443                         }
 444                         StrBufAppendBufPlain(converted_msg, ptr, tag_end - ptr, 0);
 445                         ptr = tag_end;
 446                 }
 447
 448                 /**
 449                  * Turn anything that looks like a URL into a real link, as long
 450                  * as it's not inside a tag already
 451                  */
 452                 else if ( (brak == 0) && (alevel == 0)
 453                      && (!strncasecmp(ptr, "http://", 7))) {
 454                                 /** Find the end of the link */
 455                                 int strlenptr;
 456                                 linklen = 0;
 457
 458                                 strlenptr = strlen(ptr);
 459                                 for (i=0; i<=strlenptr; ++i) {
 460                                         if ((ptr[i]==0)
 461                                            ||(isspace(ptr[i]))
 462                                            ||(ptr[i]==10)
 463                                            ||(ptr[i]==13)
 464                                            ||(ptr[i]=='(')
 465                                            ||(ptr[i]==')')
 466                                            ||(ptr[i]=='<')
 467                                            ||(ptr[i]=='>')
 468                                            ||(ptr[i]=='[')
 469                                            ||(ptr[i]==']')
 470                                            ||(ptr[i]=='"')
 471                                            ||(ptr[i]=='\'')
 472                                         ) linklen = i;
 473                                         /* did s.b. send us an entity? */
 474                                         if (ptr[i] == '&') {
 475                                                 if ((ptr[i+2] ==';') ||
 476                                                     (ptr[i+3] ==';') ||
 477                                                     (ptr[i+5] ==';') ||
 478                                                     (ptr[i+6] ==';') ||
 479                                                     (ptr[i+7] ==';'))
 480                                                         linklen = i;
 481                                         }
 482                                         if (linklen > 0) break;
 483                                 }
 484                                 if (linklen > 0) {
 485                                         char *ltreviewptr;
 486                                         char *nbspreviewptr;
 487                                         char linkedchar;
 488                                         int len;
 489
 490                                         len = linklen;
 491                                         linkedchar = ptr[len];
 492                                         ptr[len] = '\0';
 493                                         /* spot for some subject strings tinymce tends to give us. */
 494                                         ltreviewptr = strchr(ptr, '<');
 495                                         if (ltreviewptr != NULL) {
 496                                                 *ltreviewptr = '\0';
 497                                                 linklen = ltreviewptr - ptr;
 498                                         }
 499
 500                                         nbspreviewptr = strstr(ptr, "&nbsp;");
 501                                         if (nbspreviewptr != NULL) {
 502                                                 /* nbspreviewptr = '\0'; */
 503                                                 linklen = nbspreviewptr - ptr;
 504                                         }
 505                                         if (ltreviewptr != 0)
 506                                                 *ltreviewptr = '<';
 507
 508                                         ptr[len] = linkedchar;
 509
 510                                         content_length += (32 + linklen);
 511                                         StrBufAppendPrintf(converted_msg, "%s\"", new_window);
 512                                         StrBufAppendBufPlain(converted_msg, ptr, linklen, 0);
 513                                         StrBufAppendPrintf(converted_msg, "\">");
 514                                         StrBufAppendBufPlain(converted_msg, ptr, linklen, 0);
 515                                         ptr += linklen;
 516                                         StrBufAppendPrintf(converted_msg, "</A>");
 517                                 }
 518                 }
 519                 else {
 520                         StrBufAppendBufPlain(converted_msg, ptr, 1, 0);
 521                         ptr++;
 522                 }
 523
 524
 525                 if ((ptr >= msg) && (ptr <= msgend)) {
 526                         /*
 527                          * We need to know when we're inside a tag,
 528                          * so we don't turn things that look like URL's into
 529                          * links, when they're already links - or image sources.
 530                          */
 531                         if ((ptr > msg) && (*(ptr-1) == '<')) {
 532                                 ++brak;
 533                         }
 534                         if ((ptr > msg) && (*(ptr-1) == '>')) {
 535                                 --brak;
 536                                 if ((scriptlevel == 0) && (script_start_pos >= 0)) {
 537                                         StrBufCutRight(converted_msg, StrLength(converted_msg) - script_start_pos);
 538                                         script_start_pos = (-1);
 539                                 }
 540                         }
 541                         if (!strncasecmp(ptr, "</A>", 3)) --alevel;
 542                 }
 543         }
 544
 545         if (BodyArea != NULL) {
 546                 StrBufAppendBufPlain(converted_msg, HKEY("</td></tr></table>"), 0);
 547                 FreeStrBuf(&BodyArea);
 548         }
 549
 550         /**     uncomment these two lines to override conversion        */
 551         /**     memcpy(converted_msg, msg, content_length);             */
 552         /**     output_length = content_length;                         */
 553
 554         /** Output our big pile of markup */
 555         StrBufAppendBuf(Target, converted_msg, 0);
 556
 557 BAIL:   /** A little trailing vertical whitespace... */
 558         StrBufAppendPrintf(Target, "<br /><br />\n");
 559
 560         /** Now give back the memory */
 561         FreeStrBuf(&converted_msg);
 562         if ((msg != NULL) && (Source == NULL)) free(msg);
 563 }
 564
 565
 566
 567
 568
 569
 570 /*
 571  * Look for URL's embedded in a buffer and make them linkable.  We use a
 572  * target window in order to keep the Citadel session in its own window.
 573  */
 574 void UrlizeText(StrBuf* Target, StrBuf *Source, StrBuf *WrkBuf)
 575 {
 576         int len, UrlLen, Offset, TrailerLen;
 577         const char *start, *end, *pos;
 578
 579         FlushStrBuf(Target);
 580
 581         start = NULL;
 582         len = StrLength(Source);
 583         end = ChrPtr(Source) + len;
 584         for (pos = ChrPtr(Source); (pos < end) && (start == NULL); ++pos) {
 585                 if (!strncasecmp(pos, "http://", 7))
 586                         start = pos;
 587                 else if (!strncasecmp(pos, "ftp://", 6))
 588                         start = pos;
 589         }
 590
 591         if (start == NULL) {
 592                 StrBufAppendBuf(Target, Source, 0);
 593                 return;
 594         }
 595         FlushStrBuf(WrkBuf);
 596
 597         for (pos = ChrPtr(Source) + len; pos > start; --pos) {
 598                 if (  (!isprint(*pos))
 599                    || (isspace(*pos))
 600                    || (*pos == '{')
 601                    || (*pos == '}')
 602                    || (*pos == '|')
 603                    || (*pos == '\\')
 604                    || (*pos == '^')
 605                    || (*pos == '[')
 606                    || (*pos == ']')
 607                    || (*pos == '`')
 608                    || (*pos == '<')
 609                    || (*pos == '>')
 610                    || (*pos == '(')
 611                    || (*pos == ')')
 612                 ) {
 613                         end = pos;
 614                 }
 615         }
 616
 617         UrlLen = end - start;
 618         StrBufAppendBufPlain(WrkBuf, start, UrlLen, 0);
 619
 620         Offset = start - ChrPtr(Source);
 621         if (Offset != 0)
 622                 StrBufAppendBufPlain(Target, ChrPtr(Source), Offset, 0);
 623         StrBufAppendPrintf(Target, "%ca href=%c%s%c TARGET=%c%s%c%c%s%c/A%c",
 624                            LB, QU, ChrPtr(WrkBuf), QU, QU, TARGET,
 625                            QU, RB, ChrPtr(WrkBuf), LB, RB);
 626
 627         TrailerLen = StrLength(Source) - (end - ChrPtr(Source));
 628         if (TrailerLen > 0)
 629                 StrBufAppendBufPlain(Target, end, TrailerLen, 0);
 630 }
 631
 632
 633 void url(char *buf, size_t bufsize)
 634 {
 635         int len, UrlLen, Offset, TrailerLen, outpos;
 636         char *start, *end, *pos;
 637         char urlbuf[SIZ];
 638         char outbuf[SIZ];
 639
 640         start = NULL;
 641         len = strlen(buf);
 642         if (len > bufsize) {
 643                 lprintf(1, "URL: content longer than buffer!");
 644                 return;
 645         }
 646         end = buf + len;
 647         for (pos = buf; (pos < end) && (start == NULL); ++pos) {
 648                 if (!strncasecmp(pos, "http://", 7))
 649                         start = pos;
 650                 if (!strncasecmp(pos, "ftp://", 6))
 651                         start = pos;
 652         }
 653
 654         if (start == NULL)
 655                 return;
 656
 657         for (pos = buf+len; pos > start; --pos) {
 658                 if (  (!isprint(*pos))
 659                    || (isspace(*pos))
 660                    || (*pos == '{')
 661                    || (*pos == '}')
 662                    || (*pos == '|')
 663                    || (*pos == '\\')
 664                    || (*pos == '^')
 665                    || (*pos == '[')
 666                    || (*pos == ']')
 667                    || (*pos == '`')
 668                    || (*pos == '<')
 669                    || (*pos == '>')
 670                    || (*pos == '(')
 671                    || (*pos == ')')
 672                 ) {
 673                         end = pos;
 674                 }
 675         }
 676
 677         UrlLen = end - start;
 678         if (UrlLen > sizeof(urlbuf)){
 679                 lprintf(1, "URL: content longer than buffer!");
 680                 return;
 681         }
 682         memcpy(urlbuf, start, UrlLen);
 683         urlbuf[UrlLen] = '\0';
 684
 685         Offset = start - buf;
 686         if ((Offset != 0) && (Offset < sizeof(outbuf)))
 687                 memcpy(outbuf, buf, Offset);
 688         outpos = snprintf(&outbuf[Offset], sizeof(outbuf) - Offset,
 689                           "%ca href=%c%s%c TARGET=%c%s%c%c%s%c/A%c",
 690                           LB, QU, urlbuf, QU, QU, TARGET, QU, RB, urlbuf, LB, RB);
 691         if (outpos >= sizeof(outbuf) - Offset) {
 692                 lprintf(1, "URL: content longer than buffer!");
 693                 return;
 694         }
 695
 696         TrailerLen = len - (end - start);
 697         if (TrailerLen > 0)
 698                 memcpy(outbuf + Offset + outpos, end, TrailerLen);
 699         if (Offset + outpos + TrailerLen > bufsize) {
 700                 lprintf(1, "URL: content longer than buffer!");
 701                 return;
 702         }
 703         memcpy (buf, outbuf, Offset + outpos + TrailerLen);
 704         *(buf + Offset + outpos + TrailerLen) = '\0';
 705 }
 706