Opposition against committing the following?
--- snip ---
Index: ipsec_doi.c
===================================================================
RCS file: /cvsroot/src/crypto/dist/ipsec-tools/src/racoon/ipsec_doi.c,v
retrieving revision 1.36
diff -u -r1.36 ipsec_doi.c
--- ipsec_doi.c 14 Jul 2008 05:45:15 -0000 1.36
+++ ipsec_doi.c 29 Oct 2008 12:13:07 -0000
@@ -4486,7 +4486,7 @@
char *dat;
static char buf[BUFLEN];
struct ipsecdoi_id_b *id_b = (struct ipsecdoi_id_b *)id->v;
- struct sockaddr saddr;
+ struct sockaddr_storage saddr;
u_int plen = 0;
struct sockaddr isn't big enough to hold struct
sockaddr_in6, so it is (with perfect hindsight) not surprising that this
leads to trouble.
You could probably use ss_family and ss_len without so many casts, but I
suspect the way you have will be less confusing.
So looks good to me; I think you should commit it.
Attachment:
pgpBmWLxBifQ0.pgp
Description: PGP signature