Revert "serv_rssclient.c: style update"
[citadel.git] / webcit / html2html.c
diff --git a/webcit/html2html.c b/webcit/html2html.c
new file mode 100644 (file)
index 0000000..3122c75
--- /dev/null
@@ -0,0 +1,708 @@
+/*
+ * Output an HTML message, modifying it slightly to make sure it plays nice
+ * with the rest of our web framework.
+ *
+ * Copyright (c) 2005-2012 by the citadel.org team
+ *
+ * This program is open source software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 3.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include "webcit.h"
+#include "webserver.h"
+
+
+/*
+ * Strip surrounding single or double quotes from a string.
+ */
+void stripquotes(char *s)
+{
+       int len;
+
+       if (!s) return;
+
+       len = strlen(s);
+       if (len < 2) return;
+
+       if ( ( (s[0] == '\"') && (s[len-1] == '\"') ) || ( (s[0] == '\'') && (s[len-1] == '\'') ) ) {
+               s[len-1] = 0;
+               strcpy(s, &s[1]);
+       }
+}
+
+
+/*
+ * Check to see if a META tag has overridden the declared MIME character set.
+ *
+ * charset             Character set name (left unchanged if we don't do anything)
+ * meta_http_equiv     Content of the "http-equiv" portion of the META tag
+ * meta_content                Content of the "content" portion of the META tag
+ */
+void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_content)
+{
+       char *ptr;
+       char buf[64];
+
+       if (!charset) return;
+       if (!meta_http_equiv) return;
+       if (!meta_content) return;
+
+
+       if (strcasecmp(meta_http_equiv, "Content-type")) return;
+
+       ptr = strchr(meta_content, ';');
+       if (!ptr) return;
+
+       safestrncpy(buf, ++ptr, sizeof buf);
+       string_trim(buf);
+       if (!strncasecmp(buf, "charset=", 8)) {
+               strcpy(charset, &buf[8]);
+
+               /*
+                * The brain-damaged webmail program in Microsoft Exchange declares
+                * a charset of "unicode" when they really mean "UTF-8".  GNU iconv
+                * treats "unicode" as an alias for "UTF-16" so we have to manually
+                * fix this here, otherwise messages generated in Exchange webmail
+                * show up as a big pile of weird characters.
+                */
+               if (!strcasecmp(charset, "unicode")) {
+                       strcpy(charset, "UTF-8");
+               }
+
+               /* Remove wandering punctuation */
+               if ((ptr=strchr(charset, '\"'))) *ptr = 0;
+               string_trim(charset);
+       }
+}
+
+
+
+/*
+ * Sanitize and enhance an HTML message for display.
+ * Also convert weird character sets to UTF-8 if necessary.
+ * Also fixup img src="cid:..." type inline images to fetch the image
+ *
+ */
+void output_html(const char *supplied_charset, int treat_as_wiki, int msgnum, StrBuf *Source, StrBuf *Target) {
+       char buf[SIZ];
+       char *msg;
+       char *ptr;
+       char *msgstart;
+       char *msgend;
+       StrBuf *converted_msg;
+       int buffer_length = 1;
+       int line_length = 0;
+       int content_length = 0;
+       char new_window[SIZ];
+       int brak = 0;
+       int alevel = 0;
+       int scriptlevel = 0;
+       int script_start_pos = (-1);
+       int i;
+       int linklen;
+       char charset[128];
+       StrBuf *BodyArea = NULL;
+#ifdef HAVE_ICONV
+       iconv_t ic = (iconv_t)(-1) ;
+       char *ibuf;                   /* Buffer of characters to be converted */
+       char *obuf;                   /* Buffer for converted characters      */
+       size_t ibuflen;               /* Length of input buffer               */
+       size_t obuflen;               /* Length of output buffer              */
+       char *osav;                   /* Saved pointer to output buffer       */
+#endif
+       if (Target == NULL)
+               Target = WC->WBuf;
+
+       safestrncpy(charset, supplied_charset, sizeof charset);
+       msg = strdup("");
+       sprintf(new_window, "<a target=\"%s\" href=", TARGET);
+
+       if (Source == NULL) while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
+               line_length = strlen(buf);
+               buffer_length = content_length + line_length + 2;
+               ptr = realloc(msg, buffer_length);
+               if (ptr == NULL) {
+                       StrBufAppendPrintf(Target, "<b>");
+                       StrBufAppendPrintf(Target, _("realloc() error! couldn't get %d bytes: %s"),
+                                       buffer_length + 1,
+                                       strerror(errno));
+                       StrBufAppendPrintf(Target, "</b><br><br>\n");
+                       while (serv_getln(buf, sizeof buf), strcmp(buf, "000")) {
+                               /** flush */
+                       }
+                       free(msg);
+                       return;
+               }
+               msg = ptr;
+               strcpy(&msg[content_length], buf);
+               content_length += line_length;
+               strcpy(&msg[content_length], "\n");
+               content_length += 1;
+       }
+       else {
+               content_length = StrLength(Source);
+               free(msg);
+               msg = (char*) ChrPtr(Source);/* TODO: remove cast */
+               buffer_length = content_length;
+       }
+
+       /** Do a first pass to isolate the message body */
+       ptr = msg + 1;
+       msgstart = msg;
+       msgend = &msg[content_length];
+
+       while (ptr < msgend) {
+
+               /** Advance to next tag */
+               ptr = strchr(ptr, '<');
+               if ((ptr == NULL) || (ptr >= msgend)) break;
+               ++ptr;
+               if ((ptr == NULL) || (ptr >= msgend)) break;
+
+               /*
+                *  Look for META tags.  Some messages (particularly in
+                *  Asian locales) illegally declare a message's character
+                *  set in the HTML instead of in the MIME headers.  This
+                *  is wrong but we have to work around it anyway.
+                */
+               if (!strncasecmp(ptr, "META", 4)) {
+
+                       char *meta_start;
+                       char *meta_end;
+                       int meta_length;
+                       char *meta;
+                       char *meta_http_equiv;
+                       char *meta_content;
+                       char *spaceptr;
+
+                       meta_start = &ptr[4];
+                       meta_end = strchr(ptr, '>');
+                       if ((meta_end != NULL) && (meta_end <= msgend)) {
+                               meta_length = meta_end - meta_start + 1;
+                               meta = malloc(meta_length + 1);
+                               safestrncpy(meta, meta_start, meta_length);
+                               meta[meta_length] = 0;
+                               string_trim(meta);
+                               if (!strncasecmp(meta, "HTTP-EQUIV=", 11)) {
+                                       meta_http_equiv = strdup(&meta[11]);
+                                       spaceptr = strchr(meta_http_equiv, ' ');
+                                       if (spaceptr != NULL) {
+                                               *spaceptr = 0;
+                                               meta_content = strdup(++spaceptr);
+                                               if (!strncasecmp(meta_content, "content=", 8)) {
+                                                       strcpy(meta_content, &meta_content[8]);
+                                                       stripquotes(meta_http_equiv);
+                                                       stripquotes(meta_content);
+                                                       extract_charset_from_meta(charset,
+                                                                       meta_http_equiv, meta_content);
+                                               }
+                                               free(meta_content);
+                                       }
+                                       free(meta_http_equiv);
+                               }
+                               free(meta);
+                       }
+               }
+
+               /*
+                * Any of these tags cause everything up to and including
+                * the tag to be removed.
+                */     
+               if ( (!strncasecmp(ptr, "HTML", 4))
+                               ||(!strncasecmp(ptr, "HEAD", 4))
+                               ||(!strncasecmp(ptr, "/HEAD", 5))
+                               ||(!strncasecmp(ptr, "BODY", 4)) ) {
+                       char *pBody = NULL;
+
+                       if (!strncasecmp(ptr, "BODY", 4)) {
+                               pBody = ptr;
+                       }
+                       ptr = strchr(ptr, '>');
+                       if ((ptr == NULL) || (ptr >= msgend)) break;
+                       if ((pBody != NULL) && (ptr - pBody > 4)) {
+                               char* src;
+                               char *cid_start, *cid_end;
+
+                               *ptr = '\0';
+                               pBody += 4; 
+                               while ((isspace(*pBody)) && (pBody < ptr))
+                                       pBody ++;
+                               BodyArea = NewStrBufPlain(NULL,  ptr - pBody);
+
+                               if (pBody < ptr) {
+                                       src = strstr(pBody, "cid:");
+                                       if (src) {
+                                               cid_start = src + 4;
+                                               cid_end = cid_start;
+                                               while ((*cid_end != '"') && 
+                                                               !isspace(*cid_end) &&
+                                                               (cid_end < ptr))
+                                                       cid_end ++;
+
+                                               /* copy tag and attributes up to src="cid: */
+                                               StrBufAppendBufPlain(BodyArea, pBody, src - pBody, 0);
+
+                                               /* add in /webcit/mimepart/<msgno>/CID/ 
+                                                  trailing / stops dumb URL filters getting excited */
+                                               StrBufAppendPrintf(BodyArea,
+                                                               "/webcit/mimepart/%d/",msgnum);
+                                               StrBufAppendBufPlain(BodyArea, cid_start, cid_end - cid_start, 0);
+
+                                               if (ptr - cid_end > 0)
+                                                       StrBufAppendBufPlain(BodyArea, 
+                                                                       cid_end + 1, 
+                                                                       ptr - cid_end, 0);
+                                       }
+                                       else 
+                                               StrBufAppendBufPlain(BodyArea, pBody, ptr - pBody, 0);
+                               }
+                               *ptr = '>';
+                       }
+                       ++ptr;
+                       if ((ptr == NULL) || (ptr >= msgend)) break;
+                       msgstart = ptr;
+               }
+
+               /*
+                * Any of these tags cause everything including and following
+                * the tag to be removed.
+                */
+               if ( (!strncasecmp(ptr, "/HTML", 5))
+                               ||(!strncasecmp(ptr, "/BODY", 5)) ) {
+                       --ptr;
+                       msgend = ptr;
+                       strcpy(ptr, "");
+
+               }
+
+               ++ptr;
+       }
+       if (msgstart > msg) {
+               strcpy(msg, msgstart);
+       }
+
+       /* Now go through the message, parsing tags as necessary. */
+       converted_msg = NewStrBufPlain(NULL, content_length + 8192);
+
+
+       /** Convert foreign character sets to UTF-8 if necessary. */
+#ifdef HAVE_ICONV
+       if ( (strcasecmp(charset, "us-ascii"))
+                       && (strcasecmp(charset, "UTF-8"))
+                       && (strcasecmp(charset, ""))
+          ) {
+               syslog(LOG_DEBUG, "Converting %s to UTF-8\n", charset);
+               ctdl_iconv_open("UTF-8", charset, &ic);
+               if (ic == (iconv_t)(-1) ) {
+                       syslog(LOG_WARNING, "%s:%d iconv_open() failed: %s\n",
+                                       __FILE__, __LINE__, strerror(errno));
+               }
+       }
+       if  (Source == NULL) {
+               if (ic != (iconv_t)(-1) ) {
+                       ibuf = msg;
+                       ibuflen = content_length;
+                       obuflen = content_length + (content_length / 2) ;
+                       obuf = (char *) malloc(obuflen);
+                       osav = obuf;
+                       iconv(ic, &ibuf, &ibuflen, &obuf, &obuflen);
+                       content_length = content_length + (content_length / 2) - obuflen;
+                       osav[content_length] = 0;
+                       free(msg);
+                       msg = osav;
+                       iconv_close(ic);
+               }
+       }
+       else {
+               if (ic != (iconv_t)(-1) ) {
+                       StrBuf *Buf = NewStrBufPlain(NULL, StrLength(Source) + 8096);;
+                       StrBufConvert(Source, Buf, &ic);
+                       FreeStrBuf(&Buf);
+                       iconv_close(ic);
+                       msg = (char*)ChrPtr(Source); /* TODO: get rid of this. */
+               }
+       }
+
+#endif
+
+       /*
+        *      At this point, the message has been stripped down to
+        *      only the content inside the <BODY></BODY> tags, and has
+        *      been converted to UTF-8 if it was originally in a foreign
+        *      character set.  The text is also guaranteed to be null
+        *      terminated now.
+        */
+
+       if (converted_msg == NULL) {
+               StrBufAppendPrintf(Target, "Error %d: %s<br>%s:%d", errno, strerror(errno), __FILE__, __LINE__);
+               goto BAIL;
+       }
+
+       if (BodyArea != NULL) {
+               StrBufAppendBufPlain(converted_msg, HKEY("<table "), 0);  
+               StrBufAppendBuf(converted_msg, BodyArea, 0);
+               StrBufAppendBufPlain(converted_msg, HKEY(" width=\"100%\"><tr><td>"), 0);
+       }
+       ptr = msg;
+       msgend = strchr(msg, 0);
+       while (ptr < msgend) {
+
+               /** Try to sanitize the html of any rogue scripts */
+               if (!strncasecmp(ptr, "<script", 7)) {
+                       if (scriptlevel == 0) {
+                               script_start_pos = StrLength(converted_msg);
+                       }
+                       ++scriptlevel;
+               }
+               if (!strncasecmp(ptr, "</script", 8)) {
+                       --scriptlevel;
+               }
+
+               /**
+                * Change mailto: links to WebCit mail, by replacing the
+                * link with one that points back to our mail room.  Due to
+                * the way we parse URL's, it'll even handle mailto: links
+                * that have "?subject=" in them.
+                */
+               if (!strncasecmp(ptr, "<a href=\"mailto:", 16)) {
+                       content_length += 64;
+                       StrBufAppendPrintf(converted_msg,
+                                       "<a href=\"display_enter?force_room=_MAIL_?recp=");
+                       ptr = &ptr[16];
+                       ++alevel;
+                       ++brak;
+               }
+               /** Make external links open in a separate window */
+               else if (!strncasecmp(ptr, "<a href=\"", 9)) {
+                       ++alevel;
+                       ++brak;
+                       if ( ((strchr(ptr, ':') < strchr(ptr, '/')))
+                                       &&  ((strchr(ptr, '/') < strchr(ptr, '>'))) 
+                          ) {
+                               /* open external links to new window */
+                               StrBufAppendPrintf(converted_msg, new_window);
+                               ptr = &ptr[8];
+                       }
+                       else if (
+                               (treat_as_wiki)
+                               && (strncasecmp(ptr, "<a href=\"wiki?", 14))
+                               && (strncasecmp(ptr, "<a href=\"dotgoto?", 17))
+                               && (strncasecmp(ptr, "<a href=\"knrooms?", 17))
+                       ) {
+                               content_length += 64;
+                               StrBufAppendPrintf(converted_msg, "<a href=\"wiki?go=");
+                               StrBufUrlescAppend(converted_msg, WC->CurRoom.name, NULL);
+                               StrBufAppendPrintf(converted_msg, "?page=");
+                               ptr = &ptr[9];
+                       }
+                       else {
+                               StrBufAppendPrintf(converted_msg, "<a href=\"");
+                               ptr = &ptr[9];
+                       }
+               }
+               /** Fixup <img src="cid:... ...> to fetch the mime part */
+               else if (!strncasecmp(ptr, "<img ", 5)) {
+                       char *cid_start, *cid_end;
+                       char* tag_end=strchr(ptr,'>');
+                       char* src;
+                       /* FIXME - handle this situation (maybe someone opened an <img cid... 
+                        * and then ended the message)
+                        */
+                       if (!tag_end) {
+                               syslog(LOG_DEBUG, "tag_end is null and ptr is:\n");
+                               syslog(LOG_DEBUG, "%s\n", ptr);
+                               syslog(LOG_DEBUG, "Theoretical bytes remaining: %d\n", (int)(msgend - ptr));
+                       }
+
+                       src=strstr(ptr, "src=\"cid:");
+                       ++brak;
+
+                       if (src
+                           && isspace(*(src-1))
+                               && tag_end
+                               && (cid_start=strchr(src,':'))
+                               && (cid_end=strchr(cid_start,'"'))
+                               && (cid_end < tag_end)
+                       ) {
+                               /* copy tag and attributes up to src="cid: */
+                               StrBufAppendBufPlain(converted_msg, ptr, src - ptr, 0);
+                               cid_start++;
+
+                               /* add in /webcit/mimepart/<msgno>/CID/ 
+                                  trailing / stops dumb URL filters getting excited */
+                               StrBufAppendPrintf(converted_msg,
+                                               " src=\"/webcit/mimepart/%d/",msgnum);
+                               StrBufAppendBufPlain(converted_msg, cid_start, cid_end - cid_start, 0);
+                               StrBufAppendBufPlain(converted_msg, "/\"", -1, 0);
+
+                               ptr = cid_end+1;
+                       }
+                       StrBufAppendBufPlain(converted_msg, ptr, tag_end - ptr, 0);
+                       ptr = tag_end;
+               }
+
+               /**
+                * Turn anything that looks like a URL into a real link, as long
+                * as it's not inside a tag already
+                */
+               else if ( (brak == 0) && (alevel == 0) &&
+                         ( (!strncasecmp(ptr, "http://", 7)) ||
+                           (!strncasecmp(ptr, "https://", 8)))) {
+                       /** Find the end of the link */
+                       int strlenptr;
+                       linklen = 0;
+                               
+                       strlenptr = strlen(ptr);
+                       for (i=0; i<=strlenptr; ++i) {
+                               if ((ptr[i]==0)
+                                   ||(isspace(ptr[i]))
+                                   ||(ptr[i]==10)
+                                   ||(ptr[i]==13)
+                                   ||(ptr[i]=='(')
+                                   ||(ptr[i]==')')
+                                   ||(ptr[i]=='<')
+                                   ||(ptr[i]=='>')
+                                   ||(ptr[i]=='[')
+                                   ||(ptr[i]==']')
+                                   ||(ptr[i]=='"')
+                                   ||(ptr[i]=='\'')
+                                       ) linklen = i;
+                               /* did s.b. send us an entity? */
+                               if (ptr[i] == '&') {
+                                       if ((ptr[i+2] ==';') ||
+                                           (ptr[i+3] ==';') ||
+                                           (ptr[i+5] ==';') ||
+                                           (ptr[i+6] ==';') ||
+                                           (ptr[i+7] ==';'))
+                                               linklen = i;
+                               }
+                               if (linklen > 0) break;
+                       }
+                       if (linklen > 0) {
+                               char *ltreviewptr;
+                               char *nbspreviewptr;
+                               char linkedchar;
+                               int len;
+                                       
+                               len = linklen;
+                               linkedchar = ptr[len];
+                               ptr[len] = '\0';
+                               /* spot for some subject strings tinymce tends to give us. */
+                               ltreviewptr = strchr(ptr, '<');
+                               if (ltreviewptr != NULL) {
+                                       *ltreviewptr = '\0';
+                                       linklen = ltreviewptr - ptr;
+                               }
+
+                               nbspreviewptr = strstr(ptr, "&nbsp;");
+                               if (nbspreviewptr != NULL) {
+                                       /* nbspreviewptr = '\0'; */
+                                       linklen = nbspreviewptr - ptr;
+                               }
+                               if (ltreviewptr != 0)
+                                       *ltreviewptr = '<';
+
+                               ptr[len] = linkedchar;
+
+                               content_length += (32 + linklen);
+                               StrBufAppendPrintf(converted_msg, "%s\"", new_window);
+                               StrBufAppendBufPlain(converted_msg, ptr, linklen, 0);
+                               StrBufAppendPrintf(converted_msg, "\">");
+                               StrBufAppendBufPlain(converted_msg, ptr, linklen, 0);
+                               ptr += linklen;
+                               StrBufAppendPrintf(converted_msg, "</A>");
+                       }
+               }
+               else {
+                       StrBufAppendBufPlain(converted_msg, ptr, 1, 0);
+                       ptr++;
+               }
+
+
+               if ((ptr >= msg) && (ptr <= msgend)) {
+                       /*
+                        * We need to know when we're inside a tag,
+                        * so we don't turn things that look like URL's into
+                        * links, when they're already links - or image sources.
+                        */
+                       if ((ptr > msg) && (*(ptr-1) == '<')) {
+                               ++brak;
+                       }
+                       if ((ptr > msg) && (*(ptr-1) == '>')) {
+                               --brak;
+                               if ((scriptlevel == 0) && (script_start_pos >= 0)) {
+                                       StrBufCutRight(converted_msg, StrLength(converted_msg) - script_start_pos);
+                                       script_start_pos = (-1);
+                               }
+                       }
+                       if (!strncasecmp(ptr, "</A>", 3)) --alevel;
+               }
+       }
+
+       if (BodyArea != NULL) {
+               StrBufAppendBufPlain(converted_msg, HKEY("</td></tr></table>"), 0);  
+               FreeStrBuf(&BodyArea);
+       }
+
+       /**     uncomment these two lines to override conversion        */
+       /**     memcpy(converted_msg, msg, content_length);             */
+       /**     output_length = content_length;                         */
+
+       /** Output our big pile of markup */
+       StrBufAppendBuf(Target, converted_msg, 0);
+
+BAIL:  /** A little trailing vertical whitespace... */
+       StrBufAppendPrintf(Target, "<br><br>\n");
+
+       /** Now give back the memory */
+       FreeStrBuf(&converted_msg);
+       if ((msg != NULL) && (Source == NULL)) free(msg);
+}
+
+
+
+
+
+
+/*
+ * Look for URL's embedded in a buffer and make them linkable.  We use a
+ * target window in order to keep the Citadel session in its own window.
+ */
+void UrlizeText(StrBuf* Target, StrBuf *Source, StrBuf *WrkBuf)
+{
+       int len, UrlLen, Offset, TrailerLen;
+       const char *start, *end, *pos;
+       
+       FlushStrBuf(Target);
+
+       start = NULL;
+       len = StrLength(Source);
+       end = ChrPtr(Source) + len;
+       for (pos = ChrPtr(Source); (pos < end) && (start == NULL); ++pos) {
+               if (!strncasecmp(pos, "http://", 7))
+                       start = pos;
+               else if (!strncasecmp(pos, "ftp://", 6))
+                       start = pos;
+       }
+
+       if (start == NULL) {
+               StrBufAppendBuf(Target, Source, 0);
+               return;
+       }
+       FlushStrBuf(WrkBuf);
+
+       for (pos = ChrPtr(Source) + len; pos > start; --pos) {
+               if (  (!isprint(*pos))
+                  || (isspace(*pos))
+                  || (*pos == '{')
+                  || (*pos == '}')
+                  || (*pos == '|')
+                  || (*pos == '\\')
+                  || (*pos == '^')
+                  || (*pos == '[')
+                  || (*pos == ']')
+                  || (*pos == '`')
+                  || (*pos == '<')
+                  || (*pos == '>')
+                  || (*pos == '(')
+                  || (*pos == ')')
+               ) {
+                       end = pos;
+               }
+       }
+       
+       UrlLen = end - start;
+       StrBufAppendBufPlain(WrkBuf, start, UrlLen, 0);
+
+       Offset = start - ChrPtr(Source);
+       if (Offset != 0)
+               StrBufAppendBufPlain(Target, ChrPtr(Source), Offset, 0);
+       StrBufAppendPrintf(Target, "%ca href=%c%s%c TARGET=%c%s%c%c%s%c/A%c",
+                          LB, QU, ChrPtr(WrkBuf), QU, QU, TARGET, 
+                          QU, RB, ChrPtr(WrkBuf), LB, RB);
+
+       TrailerLen = StrLength(Source) - (end - ChrPtr(Source));
+       if (TrailerLen > 0)
+               StrBufAppendBufPlain(Target, end, TrailerLen, 0);
+}
+
+
+void url(char *buf, size_t bufsize)
+{
+       int len, UrlLen, Offset, TrailerLen, outpos;
+       char *start, *end, *pos;
+       char urlbuf[SIZ];
+       char outbuf[SIZ];
+
+       start = NULL;
+       len = strlen(buf);
+       if (len > bufsize) {
+               syslog(LOG_WARNING, "URL: content longer than buffer!");
+               return;
+       }
+       end = buf + len;
+       for (pos = buf; (pos < end) && (start == NULL); ++pos) {
+               if (!strncasecmp(pos, "http://", 7))
+                       start = pos;
+               if (!strncasecmp(pos, "ftp://", 6))
+                       start = pos;
+       }
+
+       if (start == NULL)
+               return;
+
+       for (pos = buf+len; pos > start; --pos) {
+               if (  (!isprint(*pos))
+                  || (isspace(*pos))
+                  || (*pos == '{')
+                  || (*pos == '}')
+                  || (*pos == '|')
+                  || (*pos == '\\')
+                  || (*pos == '^')
+                  || (*pos == '[')
+                  || (*pos == ']')
+                  || (*pos == '`')
+                  || (*pos == '<')
+                  || (*pos == '>')
+                  || (*pos == '(')
+                  || (*pos == ')')
+               ) {
+                       end = pos;
+               }
+       }
+       
+       UrlLen = end - start;
+       if (UrlLen > sizeof(urlbuf)){
+               syslog(LOG_WARNING, "URL: content longer than buffer!");
+               return;
+       }
+       memcpy(urlbuf, start, UrlLen);
+       urlbuf[UrlLen] = '\0';
+
+       Offset = start - buf;
+       if ((Offset != 0) && (Offset < sizeof(outbuf)))
+               memcpy(outbuf, buf, Offset);
+       outpos = snprintf(&outbuf[Offset], sizeof(outbuf) - Offset,  
+                         "%ca href=%c%s%c TARGET=%c%s%c%c%s%c/A%c",
+                         LB, QU, urlbuf, QU, QU, TARGET, QU, RB, urlbuf, LB, RB);
+       if (outpos >= sizeof(outbuf) - Offset) {
+               syslog(LOG_WARNING, "URL: content longer than buffer!");
+               return;
+       }
+
+       TrailerLen = len - (end - start);
+       if (TrailerLen > 0)
+               memcpy(outbuf + Offset + outpos, end, TrailerLen);
+       if (Offset + outpos + TrailerLen > bufsize) {
+               syslog(LOG_WARNING, "URL: content longer than buffer!");
+               return;
+       }
+       memcpy (buf, outbuf, Offset + outpos + TrailerLen);
+       *(buf + Offset + outpos + TrailerLen) = '\0';
+}
+