X-Git-Url: https://code.citadel.org/?a=blobdiff_plain;f=webcit%2Fhtml2html.c;h=9d6e357aaabfdf71d577063f3c5ffae85670ec14;hb=719abc348ee76436439a16bcaccce317c12d3435;hp=abd5904e846a1577fa6c2af0c5a2d3fa3e72eed6;hpb=66bc2dac8a06eac6d8fe590d1ccc867090274579;p=citadel.git

diff --git a/webcit/html2html.c b/webcit/html2html.c
index abd5904e8..9d6e357aa 100644
--- a/webcit/html2html.c
+++ b/webcit/html2html.c
@@ -8,7 +8,6 @@
  */
 /*@{*/
 #include "webcit.h"
-#include "vcard.h"
 #include "webserver.h"
 
 
@@ -59,6 +58,18 @@ void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_
 	striplt(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");
+		}
+
 	}
 }
 
@@ -67,24 +78,25 @@ void extract_charset_from_meta(char *charset, char *meta_http_equiv, char *meta_
 /**
  * \brief 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
  *
  * \param supplied_charset the input charset as declared in the MIME headers
  */
-void output_html(char *supplied_charset, int treat_as_wiki) {
+void output_html(char *supplied_charset, int treat_as_wiki, int msgnum) {
 	char buf[SIZ];
 	char *msg;
 	char *ptr;
 	char *msgstart;
 	char *msgend;
-	char *converted_msg;
-	size_t converted_alloc = 0;
+	StrBuf *converted_msg;
 	int buffer_length = 1;
 	int line_length = 0;
 	int content_length = 0;
-	int output_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];
@@ -125,7 +137,7 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
 	}
 
 	/** Do a first pass to isolate the message body */
-	ptr = msg;
+	ptr = msg + 1;
 	msgstart = msg;
 	msgend = &msg[content_length];
 
@@ -252,16 +264,27 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
 	 */
 
 	/** Now go through the message, parsing tags as necessary. */
-	converted_alloc = content_length + 8192;
-	converted_msg = malloc(converted_alloc);
+	converted_msg = NewStrBufPlain(NULL, content_length + 8192);
 	if (converted_msg == NULL) {
-		abort();			/* FIXME */
+		wprintf("Error %d: %s<br />%s:%d", errno, strerror(errno), __FILE__, __LINE__);
+		goto BAIL;
 	}
-	strcpy(converted_msg, "");
+
 	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
@@ -270,67 +293,74 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
 		 */
 		if (!strncasecmp(ptr, "<a href=\"mailto:", 16)) {
 			content_length += 64;
-			if (content_length >= converted_alloc) {
-				converted_alloc += 8192;
-				converted_msg = realloc(converted_msg, converted_alloc);
-				if (converted_msg == NULL) {
-					abort();
-				}
-			}
-			sprintf(&converted_msg[output_length],
-				"<a href=\"display_enter"
-				"?force_room=_MAIL_&recp=");
-			output_length += 47;
+			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 */
-				content_length += 64;
-				if (content_length >= converted_alloc) {
-					converted_alloc += 8192;
-					converted_msg = realloc(converted_msg, converted_alloc);
-					if (converted_msg == NULL) {
-						abort();
-					}
-				}
-				sprintf(&converted_msg[output_length], new_window);
-				output_length += strlen(new_window);
+				StrBufAppendPrintf(converted_msg, new_window);
 				ptr = &ptr[8];
 			}
 			else if ( (treat_as_wiki) && (strncasecmp(ptr, "<a href=\"wiki?", 14)) ) {
 				content_length += 64;
-				if (content_length >= converted_alloc) {
-					converted_alloc += 8192;
-					converted_msg = realloc(converted_msg, converted_alloc);
-					if (converted_msg == NULL) {
-						abort();
-					}
-				}
-				sprintf(&converted_msg[output_length], "<a href=\"wiki?page=");
-				output_length += 19;
+				StrBufAppendPrintf(converted_msg, "<a href=\"wiki?page=");
 				ptr = &ptr[9];
 			}
 			else {
-				sprintf(&converted_msg[output_length], "<a href=\"");
-				output_length += 9;
+				StrBufAppendPrintf(converted_msg, "<a href=\"");
 				ptr = &ptr[9];
 			}
 		}
+		/** Fixup <img src="cid:... ...> to fetch the mime part */
+		else if (!strncasecmp(ptr, "<img ", 5)) {
+			char* tag_end=strchr(ptr,'>');
+			char* src=strstr(ptr, " src=\"cid:");
+			char *cid_start, *cid_end;
+			++brak;
+
+			if (src && 
+				(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))) {
-				linklen = 0;
 				/** Find the end of the link */
-				for (i=0; i<=strlen(ptr); ++i) {
+				int strlenptr;
+				linklen = 0;
+				
+				strlenptr = strlen(ptr);
+				for (i=0; i<=strlenptr; ++i) {
 					if ((ptr[i]==0)
 					   ||(isspace(ptr[i]))
 					   ||(ptr[i]==10)
@@ -341,48 +371,76 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
 					   ||(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) {
-					content_length += (32 + linklen);
-					if (content_length >= converted_alloc) {
-						converted_alloc += 8192;
-						converted_msg = realloc(converted_msg, converted_alloc);
-						if (converted_msg == NULL) {
-							abort();
-						}
+					char *ltreviewptr;
+					char *nbspreviewptr;
+					char linkedchar;
+					int len = linklen;
+					
+					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;
 					}
-					sprintf(&converted_msg[output_length], new_window);
-					output_length += strlen(new_window);
-					converted_msg[output_length] = '\"';
-					converted_msg[++output_length] = 0;
-					for (i=0; i<linklen; ++i) {
-						converted_msg[output_length] = ptr[i];
-						converted_msg[++output_length] = 0;
-					}
-					sprintf(&converted_msg[output_length], "\">");
-					output_length += 2;
-					for (i=0; i<linklen; ++i) {
-						converted_msg[output_length] = *ptr++;
-						converted_msg[++output_length] = 0;
+
+					nbspreviewptr = strstr(ptr, "&nbsp;");
+					if (nbspreviewptr != NULL) {
+						///*nbspreviewptr = '\0';
+						linklen = nbspreviewptr - ptr;
 					}
-					sprintf(&converted_msg[output_length], "</A>");
-					output_length += 4;
+					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 {
-			/**
-			 * 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 == '<') ++brak;
-			if (*ptr == '>') --brak;
-			if (!strncasecmp(ptr, "</A>", 3)) --alevel;
-			converted_msg[output_length] = *ptr++;
-			converted_msg[++output_length] = 0;
+			StrBufAppendBufPlain(converted_msg, ptr, 1, 0);
+			ptr++;
+		}
+
+		/**
+		 * 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-1) == '<') {
+			++brak;
+		}
+		if (*(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;
 	}
 
 	/**	uncomment these two lines to override conversion	*/
@@ -390,13 +448,13 @@ void output_html(char *supplied_charset, int treat_as_wiki) {
 	/**	output_length = content_length;				*/
 
 	/** Output our big pile of markup */
-	client_write(converted_msg, output_length);
+	StrBufAppendBuf(WC->WBuf, converted_msg, 0);
 
-	/** A little trailing vertical whitespace... */
+BAIL:	/** A little trailing vertical whitespace... */
 	wprintf("<br /><br />\n");
 
 	/** Now give back the memory */
-	if (converted_msg != NULL) free(converted_msg);
+	FreeStrBuf(&converted_msg);
 	if (msg != NULL) free(msg);
 }