Major Bugfix: Buffer overrun in renewAddress

It was right there the whole time! This painful bug took forever to
find, and it sooo shouldn't have...  sizeof(contactNode) is 8, not 2...

This bug has existed since
590d4d6
on Dec 1, 2015 and caused all kinds of mayhem, since it was overwriting
a memory space that changes with changes to the code, so symptoms seemed
vary significantly.

On RPi, segfaults and odd behavior were the main issues. This will
affect all platforms, but only on RPi was it really noticeable so far.

RF24Mesh.cpp

@@ -284,7 +284,7 @@ bool RF24Mesh::requestAddress(uint8_t level){
 		bool goodSignal = radio.testRPD();
         #endif
 		if(network.update() == NETWORK_POLL){
-            memcpy(&contactNode[pollCount],&network.frame_buffer[0],sizeof(contactNode));
+            memcpy(&contactNode[pollCount],&network.frame_buffer[0],sizeof(uint16_t));
             ++pollCount;
 
             #if defined (MESH_DEBUG_SERIAL) || defined (MESH_DEBUG_PRINTF)    
@@ -460,10 +460,10 @@ void RF24Mesh::setAddress(uint8_t nodeID, uint16_t address){
   }
 
    #if defined (__linux)  && !defined(__ARDUINO_X86__)
-		if(millis()-lastFileSave > 300){
-			lastFileSave = millis();
+		//if(millis()-lastFileSave > 300){
+		//	lastFileSave = millis();
 			saveDHCP();
-		}
+		//}
    #endif	  
 
 }