From: Art Cancro <ajc@citadel.org>
Date: Fri, 22 Jul 2005 04:03:31 +0000 (+0000)
Subject: * HTML messages in foreign character sets are now converted to UTF-8 for
X-Git-Tag: v7.86~4778
X-Git-Url: https://code.citadel.org/?p=citadel.git;a=commitdiff_plain;h=f9f392b97bb7b510c25dcd72db49c747715784b9

* HTML messages in foreign character sets are now converted to UTF-8 for
  display.
---

diff --git a/webcit/ChangeLog b/webcit/ChangeLog
index 125fb66b7..2830db543 100644
--- a/webcit/ChangeLog
+++ b/webcit/ChangeLog
@@ -1,4 +1,8 @@
 $Log$
+Revision 619.30  2005/07/22 04:03:30  ajc
+* HTML messages in foreign character sets are now converted to UTF-8 for
+  display.
+
 Revision 619.29  2005/07/22 03:40:22  ajc
 * Handle display of RFC2047-encoded subjects and senders by converting
   them to UTF-8 (all WebCit pages are output as UTF-8).  Resolves bug #136.
@@ -2728,4 +2732,3 @@ Sun Dec  6 19:50:55 EST 1998 Art Cancro <ajc@uncnsrd.mt-kisco.ny.us>
 
 1998-12-03 Nathan Bryant <bryant@cs.usm.maine.edu>
 	* webserver.c: warning fix
-
diff --git a/webcit/html2html.c b/webcit/html2html.c
index 271dbe143..110d9682e 100644
--- a/webcit/html2html.c
+++ b/webcit/html2html.c
@@ -24,14 +24,21 @@
 #include <stdarg.h>
 #include <pthread.h>
 #include <signal.h>
+
+#ifdef HAVE_ICONV
+#include <iconv.h>
+#endif
+
 #include "webcit.h"
 #include "vcard.h"
 #include "webserver.h"
 
 
 /*
+ * Sanitize and enhance an HTML message for display.
+ * Also convert weird character sets to UTF-8 if necessary.
  */
-void output_html(void) {
+void output_html(char *charset) {
 	char buf[SIZ];
 	char *msg;
 	char *ptr;
@@ -47,6 +54,14 @@ void output_html(void) {
 	int alevel = 0;
 	int i;
 	int linklen;
+#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
 
 	msg = strdup("");
 	sprintf(new_window, "<A TARGET=\"%s\" HREF=", TARGET);
@@ -68,6 +83,29 @@ void output_html(void) {
 		content_length += 1;
 	}
 
+#ifdef HAVE_ICONV
+	if ( (strcasecmp(charset, "us-ascii"))
+	   && (strcasecmp(charset, "UTF-8")) ) {
+		ic = iconv_open("UTF-8", charset);
+		if (ic == (iconv_t)(-1) ) {
+			lprintf(5, "iconv_open() failed: %s\n", strerror(errno));
+		}
+	}
+	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);
+	}
+#endif
+
 	ptr = msg;
 	msgstart = msg;
 	msgend = &msg[content_length];
diff --git a/webcit/messages.c b/webcit/messages.c
index c20ebc455..370454c79 100644
--- a/webcit/messages.c
+++ b/webcit/messages.c
@@ -784,7 +784,7 @@ void read_message(long msgnum) {
 
 	else /* HTML is fun, but we've got to strip it first */
 	if (!strcasecmp(mime_content_type, "text/html")) {
-		output_html();
+		output_html(mime_charset);
 	}
 
 	/* Unknown weirdness */
diff --git a/webcit/webcit.h b/webcit/webcit.h
index 699854449..5fc1b2b56 100644
--- a/webcit/webcit.h
+++ b/webcit/webcit.h
@@ -429,7 +429,7 @@ void display_addressbook(long msgnum, char alpha);
 void offer_start_page(void);
 void convenience_page(char *titlebarcolor, char *titlebarmsg, char *messagetext);
 void change_start_page(void);
-void output_html(void);
+void output_html(char *);
 void display_floorconfig(char *);
 void delete_floor(void);
 void create_floor(void);