fix: attempt to handle assertion error due to ungraceful shutdown on …

…client restart

Replace assertion with proper error handling in checkpointsync.cpp to gracefully handle cases where the client doesn't shut down properly. 

This change:

- Handles the assertion that occasionally causes the client to become unresponsive
- Implements a multi-step recovery process to find valid checkpoints
- Adds detailed logging for diagnostic purposes
- Maintains chain consistency by preferring recent valid blocks
- Suggests a reindex as a last resort

This fix prevents the client from crashing when mapBlockIndex doesn't contain the expected checkpoint after an ungraceful shutdown (power loss, hardware failure, etc).

src/checkpointsync.cpp

@@ -210,12 +210,30 @@ bool CheckSyncCheckpoint(const uint256 hashBlock, const int nHeight, const CBloc
             return true;
         }
     }
-
+    
     const CBlockIndex* pindexSync;
     {
         LOCK2(cs_main, cs_hashSyncCheckpoint);
-        // sync-checkpoint should always be accepted block
-        assert(mapBlockIndex.count(hashSyncCheckpoint));
+        // sync-checkpoint should be an accepted block
+        if (!mapBlockIndex.count(hashSyncCheckpoint)) {
+            LogPrintf("Checkpoints: %s: WARNING - Sync checkpoint %s not found in block index after unclean shutdown\n", 
+                     __func__, hashSyncCheckpoint.ToString());
+            LogPrint(BCLog::NET, "Checkpoints: Attempting recovery of sync checkpoint state\n");
+
+            // Attempt to recover by resetting the sync checkpoint
+            if (!ResetSyncCheckpoint()) {
+                // If reset fails, we need to inform the user that a reindex might be necessary
+                LogPrintf("Checkpoints: %s: CRITICAL - Failed to reset sync checkpoint. A reindex may be required.\n", __func__);
+                return error("%s: Sync checkpoint recovery failed - please restart with -reindex", __func__);
+            }
+
+            // After reset, we should have a valid checkpoint
+            if (!mapBlockIndex.count(hashSyncCheckpoint)) {
+                LogPrintf("Checkpoints: %s: CRITICAL - Failed to recover sync checkpoint state after reset attempt\n", __func__);
+                return error("%s: Failed to recover sync checkpoint state", __func__);
+            }
+            LogPrint(BCLog::NET, "Checkpoints: Successfully recovered sync checkpoint state\n");
+        }
         pindexSync = mapBlockIndex[hashSyncCheckpoint];
     }