Make LRU implementation bi-directional

Expanded the LRU node and queue structures to store both UID and name
and manage them through dedicated UID and name hash tables.

Introduced dual-lookup routines for retrieving cache entries by UID or
name and wired them into the initialization and teardown logic.

Updated UID/GID lookup helpers and normalization to leverage the new
bidirectional LRU cache instead of direct system queries.

Introduced tests verifying basic cache hits/misses and bidirectional
UID↔name mapping to ensure consistent behavior of the new LRU
implementation.

Author: stevegrubb

auparse/internal.h

@@ -192,6 +192,8 @@ int aup_load_config(auparse_state_t *au, struct daemon_conf *config,
 		    log_test_t lt);
 void aup_free_config(struct daemon_conf *config);
 
+/* Resolve @name to a uid, caching the result for future lookups. */
+uid_t lookup_uid_from_name(auparse_state_t *au, const char *name);
 // normalize.c
 void init_normalizer(normalize_data *d);
 void clear_normalizer(normalize_data *d);