react-native 0.10 update > NSNumber about react-native-fbsdk

[amccloud]

I'm seeing the same error. Looks like this was added by facebook/react-native@95b9dd3

I've checked the call-site and connectionId is 0. I've also tried making connectionId start from 1.

[amccloud]

I completely looked over the other arguments. The issue is that the timeout argument must be set now.

[amccloud]

It appears that the callback also can't be null. Work around is to dispatch your request manually instead of calling send() on it.

FBSDKGraphRequestManager.batchRequests([profileRequest], function() {}, 60);

[tnguyen42]

Should we wait for a fix or should I manually patch it for the moment? ^^

[mponizil]

FBSDKGraphRequestManager.batchRequests([profileRequest], function() {}, 60);

suppresses the red error screen, but still getting The operation couldn’t be completed in Chrome debugger. My code:

var fetchFriendsRequest = new FBSDKGraphRequest((error, result) => {
  console.log('FBSDKGraphRequest', error, result);
}, '/me/friends?fields=id,name');

User has been authenticated and authorized user_friends read permissions.

[email protected]
[email protected]

Edit: sorry -- I think this was from not having FB callback in AppDelegate.

[ceditoph]

Doesn't null evaluate to 0?
Where is the connectionID being set?

[amccloud]

@ceditoph Sorry, connectionID is correct. The issue is the timeout and callback are null.

[ceditoph]

@amCloud thanks. Where are you assigning profileRequest? I don't see an example.

[kent10ou]

Hey I'm working with Ceditoph. We're diving into the source code and can't figure out the proper args to pass in

constructor(callback: (error: ?Object, result: ?Object) => void, graphPath: string, parameters: ?Object, tokenString: ?string, version: ?string, HTTPMethod: ?string) {
this.callback = callback;
this.graphPath = graphPath;
this.parameters = parameters ? parameters : {};
this.tokenString = tokenString;
this.version = version;
this.HTTPMethod = HTTPMethod;
this.graphErrorRecoveryDisabled = false;
}

[amccloud]

@ceditoph it is something like this

var profileRequest = new FBSDKGraphRequest((error, result) => {
  if (error) {
    alert('Error making request.');
  } else {
    // Data from request is in result
  }
}, '/me');

[ceditoph]

@amccloud thanks! it's working. I don't really like how they have it set up, but the timeout is a solid workaround!
@kent10ou and I will keep at it.

[ceditoph]

@amccloud so we have it working without getting an error, but how do I grab the result of batch running profileRequest?
Can I use a promise, like .then()?

Thanks again

[ceditoph]

@amccloud nvm. I see the result from the output of fetchFriends

[khanetor]

It's been a while, is this considered a bug?

[valamelkor]

Yep. This is a bug from my point of view. It prevented us from upgrading from 0.9.

The actual working piece of code:

FBHelper.getFriends = function (success, fail)
{
  var fetchFriends = new FBSDKGraphRequest((error, result) => {
    if (error)
    {
      console.log("->FBHelper.getFriends Error: ", error);
      fail(error);
    }
    else
    {
        success(result);
    }
  }, '/me/friends?fields=id,name,picture' 
  );

  //fetchFriends.start(); //works before 0.10
  FBSDKGraphRequestManager.batchRequests([fetchFriends], function() {}, 60); // works for 0.10 and later
};

Thanks @amccloud for workaround, but for me it took some time to figure out how it should be implemented. :)

[weeweewee]

New to react-native and fbsdk here but after looking at the source code, assigning a timeout to .start(60) seemed to be a workaround as well. However, RCTFBSDKGraphRequestManager doesn't seem to like the fact that the callback is null.
A simple fix to GraphRequest.ios.js/start(timeout: number) where changing the callback from null to function() {} allows us to use the start function once again.
Feels hackish and I'm not too confident about this so I'm leaving it as a comment here.

[shaneosullivan]

Any movement on getting this properly fixed without hacks? The SDK is well behind RN in general, working with v0.8.

[dzhuowen]

This issue has been handled by #44