fix: Ember: revert default APS options change in #1065 (#1068)

src/adapter/ember/adapter/emberAdapter.ts

@@ -368,11 +368,14 @@ const STACK_CONFIGS = {
 };
 
 /**
- * NOTE: This from SDK:
+ * NOTE: This from SDK is currently ignored here because of issues in below links:
  * - BUGZID 12261: Concentrators use MTORRs for route discovery and should not enable route discovery in the APS options.
- * - https://community.silabs.com/s/question/0D58Y00008DRfDCSA1/coordinator-cant-send-unicast-to-sleepy-node-after-reboot?language=en_US
+ * - https://community.silabs.com/s/question/0D58Y00008DRfDCSA1/coordinator-cant-send-unicast-to-sleepy-node-after-reboot
+ * - https://community.silabs.com/s/question/0D58Y0000B4nTb7SQE/largedense-network-communication-problem-source-route-table-not-big-enough
+ * 
+ * Removing `ENABLE_ROUTE_DISCOVERY` leads to devices that won't reconnect/go offline, and various other issues. Keeping it for now.
  */
-const DEFAULT_APS_OPTIONS = (EmberApsOption.RETRY | EmberApsOption.ENABLE_ADDRESS_DISCOVERY);
+const DEFAULT_APS_OPTIONS = (EmberApsOption.RETRY | EmberApsOption.ENABLE_ROUTE_DISCOVERY | EmberApsOption.ENABLE_ADDRESS_DISCOVERY);
 /**
  * Enabling this allows to immediately reject requests that won't be able to get to their destination.
  * However, it causes more NCP calls, notably to get the source route overhead.

src/adapter/ember/ezsp/ezsp.ts

@@ -2378,7 +2378,7 @@ export class Ezsp extends EventEmitter {
      * @param lastHopRssi int8_t The energy level (in units of dBm) observed during the reception.
      */
     ezspNetworkFoundHandler(networkFound: EmberZigbeeNetwork, lastHopLqi: number, lastHopRssi: number): void {
-        logger.debug(`ezspNetworkFoundHandler(): callback called with: [networkFound=${networkFound}], `
+        logger.debug(`ezspNetworkFoundHandler(): callback called with: [networkFound=${JSON.stringify(networkFound)}], `
             + `[lastHopLqi=${lastHopLqi}], [lastHopRssi=${lastHopRssi}]`, NS);
     }