From: Art Cancro <ajc@citadel.org>
Date: Sat, 24 Feb 2007 04:58:54 +0000 (+0000)
Subject: Updated the html2html filter.  When we encounter a charset
X-Git-Tag: v7.86~3558
X-Git-Url: https://code.citadel.org/?p=citadel.git;a=commitdiff_plain;h=4c6e691894791003cc46faa150674bf2f8a11c6d

Updated the html2html filter.  When we encounter a charset
declaration within the meta tags, the charset 'unicode' is manually aliased
to 'UTF-8'.  GNU iconv wants to treat it as an alias for 'UTF-16' but
messages generated by brain-damaged MS Exchange webmail are declaring
'unicode' when they really mean 'UTF-8'.  This change only applies to
meta tags; it is not a program-wide alias.
---

diff --git a/webcit/html2html.c b/webcit/html2html.c
index 8799aaa2a..35fffd5d2 100644
--- a/webcit/html2html.c
+++ b/webcit/html2html.c
@@ -59,6 +59,18 @@ 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");
+		}
+
 	}
 }