find is uninitialized and only has a small portion of find.addr.inet
set in fr_redis_cluster_pool_by_node_addr(), but then all of find.addr
is assigned to spare->pending_addr, hence coverity complains. It turna
out, though, that cluster_node_connect() only uses the part that did
get set.
pthread_mutex_unlock(&cluster->mutex);
return -1;
}
+ /* coverity[uninit_use] */
spare->pending_addr = find.addr; /* Set the config to be applied */
if (cluster_node_connect(cluster, spare) < 0) {
pthread_mutex_unlock(&cluster->mutex);