Re: [Pkg-net-snmp-devel] Bug#498475: Lenny-Ignore tag request for SNMP related bugs
Hi Andreas,
> * Vincent Bernat (bernat@debian.org) [080910 12:08]:
>> Those packages ship MIB files that are non-free (same license as IETF
>> RFC). However, there is no quick fix: removing those files make the
>> packages unusable, moving to non-free make a lot of important package
>> (snmpd) move to contrib, etc.
>
> I fear you're right. :(
Thanks.
I'd like to request an unblock for net-snmp to include two fixes (in addition to
translation updates):
- patch 52_suppress_registration_warnings.patch: Official Patch 1805971.
This patch suppresses certain syslog messages upstream believes to be harmless:
Index: agent/agent_registry.c
===================================================================
--- net-snmp-5.4.1.orig/agent/agent_registry.c (Revision 16668)
+++ net-snmp-5.4.1/agent/agent_registry.c (Revision 16669)
@@ -532,7 +532,8 @@
if (next && (next->namelen == new_sub->namelen) &&
(next->priority == new_sub->priority)) {
- netsnmp_assert(!"registration != duplicate"); /* always false */
+ if (new_sub->namelen != 1) /* ignore root OID dups */
+ netsnmp_assert(!"registration != duplicate"); /* always false */
return MIB_DUPLICATE_REGISTRATION;
}
- patch 51_allow_g_groupname.patch adds the possibility to use a group name in additon
to a group id. As this capability has been requested three times, I'd like to add
this patch to Lenny. The patch has been rewiewed and included by upstream:
Index: agent/snmpd.c
===================================================================
--- net-snmp-5.4.1.orig/agent/snmpd.c (Revision 17157)
+++ net-snmp-5.4.1/agent/snmpd.c (Revision 17158)
@@ -574,8 +574,26 @@
#if HAVE_UNISTD_H
case 'g':
if (optarg != NULL) {
+ char *ecp;
+ int gid;
+
+ gid = strtoul(optarg, &ecp, 10);
+ if (*ecp) {
+#if HAVE_GETPWNAM && HAVE_PWD_H
+ struct group *info;
+ info = getgrnam(optarg);
+ if (info) {
+ gid = info->gr_gid;
+ } else {
+#endif
+ fprintf(stderr, "Bad group id: %s\n", optarg);
+ exit(1);
+#if HAVE_GETPWNAM && HAVE_PWD_H
+ }
+#endif
+ }
netsnmp_ds_set_int(NETSNMP_DS_APPLICATION_ID,
- NETSNMP_DS_AGENT_GROUPID, atoi(optarg));
+ NETSNMP_DS_AGENT_GROUPID, gid);
} else {
usage(argv[0]);
}
I have an additional patch pending in svn which fixes a regression introduced by a
fix for CVE-2008-2292, however I'd like to wait for the decission of these patches first.
Thanks,
Jochen
Reply to: