Occasional crash when inviting peer in multipeer connectivity

Question

I have a published app that uses Multipeer conncectivity to send small amounts of data at regular intervals. I am using MCNearbyServiceAdvertiser and MCNearbyServiceBrowser. Everything works fine when I test. However, I have been receiving crash reports from my users, not a lot, but enough to deserve my attention.

*** Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '*** -   [__NSPlaceholderDictionary initWithObjects:forKeys:count:]: attempt to insert nil object from objects[2]'

Last Exception Backtrace:
0   CoreFoundation                       0x30b8cf03 __exceptionPreprocess + 131
1   libobjc.A.dylib                      0x3b321ce7 objc_exception_throw + 36
2   CoreFoundation                       0x30acad3f -[__NSPlaceholderDictionary initWithObjects:forKeys:count:] + 532
3   CoreFoundation                       0x30acab03 +[NSDictionary dictionaryWithObjects:forKeys:count:] + 48
4   MultipeerConnectivity                0x32420e87 -[MCNearbyServiceBrowser syncInvitePeer:toSession:withContext:timeout:] + 452
5   MultipeerConnectivity                0x324220cf __67-[MCNearbyServiceBrowser invitePeer:toSession:withContext:timeout:]_block_invoke + 84
6   libdispatch.dylib                    0x3b80ad53 _dispatch_call_block_and_release + 8
7   libdispatch.dylib                    0x3b80fcbd _dispatch_queue_drain + 486
8   libdispatch.dylib                    0x3b80cc6f _dispatch_queue_invoke + 40
9   libdispatch.dylib                    0x3b8105f1 _dispatch_root_queue_drain + 74
10  libdispatch.dylib                    0x3b8108dd _dispatch_worker_thread2 + 54
11  libsystem_pthread.dylib              0x3b93bc17 _pthread_wqthread + 296
12  libsystem_pthread.dylib              0x3b93badc start_wqthread + 6

Despite a lot of efforts I have been unable to recreate the crash. I have tried putting the app in the background, loosing connection, turning the network on and off etc. Usually everthing reconnects nicely, sometimes after tapping a "reconnect button", but no crashes. Does anyone have a tip about the cause or at least how to force this o happen?

The browser side looks like this:

-(void)setUpPP{
  self.PPid = [[MCPeerID alloc] initWithDisplayName:@"PhotoFinish"];
  self.SSid = [[MCPeerID alloc] init];

  self.PPsession = [[MCSession alloc] initWithPeer:self.PPid securityIdentity:nil encryptionPreference:MCEncryptionNone];
  self.PPsession.delegate = self;

  self.PPbrowser = [[MCNearbyServiceBrowser alloc] initWithPeer:self.PPid serviceType:@"sprinttimer" ];
  self.PPbrowser.delegate = self;
  [self.PPbrowser startBrowsingForPeers];
}

-(IBAction)startSync:(id)sender {
  [self.PPbrowser invitePeer:self.SSid toSession:self.PPsession withContext:nil timeout:10];
}

-(void)session:(MCSession *)session peer:(MCPeerID *)peerID didChangeState:(MCSessionState)state{

  NSString *displayString;
  if (state == MCSessionStateConnected && self.PPsession) {
    displayString=@"Start Sender connected";

  } else if (state == MCSessionStateNotConnected && self.PPsession) {
    displayString=@"No Start Sender in range";
    connected=NO;
  }
 [self performSelectorOnMainThread:@selector(displaySync:) withObject:displayString waitUntilDone:NO];
}

-(void)browser:(MCNearbyServiceBrowser *)browser foundPeer:(MCPeerID *)peerID withDiscoveryInfo:(NSDictionary *)info{
  self.SSid=peerID;
}

Answer 1

It would be helpful to know what action triggers startSync: but in setUpPP you are assigning a meaningless instance of MCPeerID (with no displayName) to self.SSid so I could see that crash happening if startSync: is ever called before self.SSid gets a real MCPeerID assigned to it.

My suggestion would be to start out with self.SSid as nil, and make sure that you are not sending invitations unless that property has a value. Going further I wouldn't rely on a single property to store an MCPeerID that could have it's value changed before startSync: is called.