See ChangeLog: Sun Jan 24 18:16:26 CET 1999 Werner Koch

1 files changed, 15 insertions, 2 deletions

diff --git a/g10/parse-packet.c b/g10/parse-packet.c
index 53d8c7bfb..3c45a45b4 100644
--- a/g10/parse-packet.c
+++ b/g10/parse-packet.c
@@ -1397,15 +1397,28 @@ parse_user_id( IOBUF inp, int pkttype, unsigned long pktlen, PACKET *packet )
 {
     byte *p;
 
-    packet->pkt.user_id = m_alloc(sizeof *packet->pkt.user_id  + pktlen - 1);
+    packet->pkt.user_id = m_alloc(sizeof *packet->pkt.user_id  + pktlen);
     packet->pkt.user_id->len = pktlen;
     p = packet->pkt.user_id->name;
-    for( ; pktlen; pktlen--, p++ )
+    for( ; pktlen; pktlen--, p++ ) {
 	*p = iobuf_get_noeof(inp);
+	/* 0xff is not a valid utf-8 encoding so we can use it to replace
+	 * Nulls. This has the advantage that we can work with regular
+	 * C strings. When exporting it, we change it back to Null
+	 * the utf-8 functions know about this special convention.
+	 * The advantage of this single character is that we can
+	 * simple replace it.  Problem is that we can't handle the 0xff
+	 * character which may have been used by pref rfc2440 implementations
+	 * I hope we can live with this. */
+	if( !*p )
+	    *p = 0xff;
+    }
+    *p = 0;
 
     if( list_mode ) {
 	int n = packet->pkt.user_id->len;
 	printf(":user id packet: \"");
+	/* fixme: Hey why don't we replace this wioth print_string?? */
 	for(p=packet->pkt.user_id->name; n; p++, n-- ) {
 	    if( *p >= ' ' && *p <= 'z' )
 		putchar(*p);