feat(ha-repl-hardening): safe_committed_lsn from vote replies (task 4.1)

- Add ha_compute_safe_committed_lsn() in ha.c: collects LSNs from
  granted voters (self + peers), sorts descending, picks index
  floor(n_granted/2) for majority-threshold value
- Wire into raft_cb_become_leader: compute safe_lsn/safe_tag from
  ha->elect_ctx after winning election, call persist_truncate_wal_above
- Four unit tests in test_ha_evaluate_committing.c:
  voters [99,100]+self=100 → 100; voters [50,51]+self=100 → 51;
  single-node → 0; denied voter excluded → self-only

src/controller/ha.h

@@ -289,6 +289,22 @@ int ha_apply_wal_catchup_entry(ha_state_t *ha, persist_t *p,
 void ha_nop_reply_state_report(ha_state_t *ha, myconf_t *myconf,
                                 state_report_t *sr, xce_reply_t *reply_out);
 
+/* ── Task 4.1 (ha-repl-hardening): safe commit point computation ─────────────
+ *
+ * ha_compute_safe_committed_lsn: determine safe_committed_lsn and safe_config_tag
+ * from vote replies collected in ctx after winning an election.
+ *
+ * Collects LSNs from all granted voters (self + peers with vote_granted=1),
+ * sorts descending, and returns the majority-threshold element.
+ *
+ * single-node (n_peers=0): outputs 0 for both fields (no vote data available).
+ *
+ * Requirements: 2.3 (ha-repl-hardening spec)
+ */
+void ha_compute_safe_committed_lsn(ha_elect_ctx_t *ctx,
+                                    uint64_t       *safe_lsn_out,
+                                    uint64_t       *safe_tag_out);
+
 /* ── Task 3.2: Unified replication decision enum and evaluator ────────────── */
 
 /*