Meteor: Proper use of Meteor.wrapAsync on server

From the Meteor.wrapAsync http://docs.meteor.com/#meteor_wrapasync you can see that you need to pass it a function and optionally a context, whereas on your two attempts you are passing the RESULT of calling the async version of Stripe.customers.create.

Meteor.methods({
  stripeCreateUser: function(options) {
    // get a sync version of our API async func
    var stripeCustomersCreateSync=Meteor.wrapAsync(Stripe.customers.create,Stripe.customers);
    // call the sync version of our API func with the parameters from the method call
    var result=stripeCustomersCreateSync({
      description: 'Woot! A new customer!',
      card: options.ccToken,
      plan: options.pricingPlan
    });
    // do whatever you want with the result
    console.log(result);
  }
});

Meteor.wrapAsync transforms an async function into a convenient synchronous-looking function which allows to write sequentially looking code. (underhood everything is still executed within the asynchronous Node.js event loop).

We need to pass to Meteor.wrapAsync our API function (Stripe.customers.create) along with the function context, ie this inside the body of the API func, which in this case is Stripe.customers.

EDIT :

How to retrieve errors ?

Traditional node style API functions usually take a callback as last argument that will get called ultimately when the required task is completed. This callback takes 2 arguments : error and data, either one will be null according to the result of the call.

How do we access the error object using the synchronous wrapped functions returned by Meteor.wrapAsync ?

We have to rely on using try/catch blocks, because in case of error, it will be thrown by the sync function instead of being passed as first argument of the async function callback.

try{
  var result=syncFunction(params);
  console.log("result :",result);
}
catch(error){
  console.log("error",error);
}
// is the equivalent of :
asyncFunc(params,function(error,result){
  if(error){
    console.log("error",error);
    return;
  }
  console.log("result :",result);
});

why doesn't Stripe need to be passed?

JavaScript has no "namespace" concept, so API developers use the common trick of defining a global object acting as API namespace, properties defined on this object are "sub-modules" of the API. It means that Stripe.customers is a submodule of the Stripe API exposing customers-related funcs, and as such these funcs this context is Stripe.customers, not Stripe.

You can test it yourself by copy pasting this mocking code in your browser console :

Stripe={
  customers:{
    create:function(){
      console.log(this==Stripe.customers);
    }
  }
};

And then calling the stub function in your browser console like this :

> Stripe.customers.create();
true

Another option is this package which achieves the similar goals.

meteor add meteorhacks:async

From the package README:

Async.wrap(function)

Wrap an asynchronous function and allow it to be run inside Meteor without callbacks.

//declare a simple async function
function delayedMessge(delay, message, callback) {
  setTimeout(function() {
    callback(null, message);
  }, delay);
}

//wrapping
var wrappedDelayedMessage = Async.wrap(delayedMessge);

//usage
Meteor.methods({
  'delayedEcho': function(message) {
    var response = wrappedDelayedMessage(500, message);
    return response;
  }
});