mpromise

Promises A+ conformant implementation

#mpromise

A promises/A+ conformant implementation, written for mongoose.

$ npm install mpromise

An mpromise can be in any of three states, pending, fulfilled (success), or rejected (error). Once it is either fulfilled or rejected it's state can no longer be changed.

The exports object is the Promise constructor.

var Promise = require('mpromise');

The constructor accepts an optional function which is executed when the promise is first resolved (either fulfilled or rejected).

var promise = new Promise(fn);

This is the same as passing the fn to onResolve directly.

var promise = new Promise;
promise.onResolve(function (errargs..) {
  ...
});

####fulfill

Fulfilling a promise with values:

var promise = new Promise;
promise.fulfill(args...);

If the promise has already been fulfilled or rejected, no action is taken.

####reject

Rejecting a promise with a reason:

var promise = new Promise;
promise.reject(reason);

If the promise has already been fulfilled or rejected, no action is taken.

####resolve

Node.js callback style promise resolution (err, args...):

var promise = new Promise;
promise.resolve([reason], [arg1, arg2, ...]);

If the promise has already been fulfilled or rejected, no action is taken.

####onFulfill

To register a function for execution when the promise is fulfilled, pass it to onFulfill. When executed it will receive the arguments passed to fulfill().

var promise = new Promise;
promise.onFulfill(function (ab) {
  assert.equal(3, a + b);
});
promise.fulfill(1, 2);

The function will only be called once when the promise is fulfilled, never when rejected.

Registering a function with onFulfill after the promise has already been fulfilled results in the immediate execution of the function with the original arguments used to fulfill the promise.

var promise = new Promise;
promise.fulfill(" :D ");
promise.onFulfill(function (arg) {
  console.log(arg); // logs " :D " 
})

####onReject

To register a function for execution when the promise is rejected, pass it to onReject. When executed it will receive the argument passed to reject().

var promise = new Promise;
promise.onReject(function (reason) {
  assert.equal('sad', reason);
});
promise.reject('sad');

The function will only be called once when the promise is rejected, never when fulfilled.

Registering a function with onReject after the promise has already been rejected results in the immediate execution of the function with the original argument used to reject the promise.

var promise = new Promise;
promise.reject(" :( ");
promise.onReject(function (reason) {
  console.log(reason); // logs " :( " 
})

####onResolve

Allows registration of node.js style callbacks (err, args..) to handle either promise resolution type (fulfill or reject).

// fulfillment 
var promise = new Promise;
promise.onResolve(function (errab) {
  console.log(+ b); // logs 3 
});
promise.fulfill(1, 2);
 
// rejection 
var promise = new Promise;
promise.onResolve(function (err) {
  if (err) {
    console.log(err.message); // logs "failed" 
  }
});
promise.reject(new Error('failed'));

####then

Creates a new promise and returns it. If onFulfill or onReject are passed, they are added as SUCCESS/ERROR callbacks to this promise after the nextTick.

Conforms to promises/A+ specification and passes its tests.

// promise.then(onFulfill, onReject); 
 
var p = new Promise;
 
p.then(function (arg) {
  return arg + 1;
}).then(function (arg) {
  throw new Error(arg + ' is an error!');
}).then(null, function (err) {
  assert.ok(err instanceof Error);
  assert.equal('2 is an error', err.message);
});
p.fullfill(1);

####end

Signifies that this promise was the last in a chain of then()s: if a handler passed to the call to then which produced this promise throws, the exception be rethrown. You can pass an OnReject handler to end so that exceptions will be handled (like a final catch clause); This method returns it's promise for easy use with return.

var p = new Promise;
p.then(function(){ throw new Error('shucks') });
setTimeout(function () {
  p.fulfill();
  // error was caught and swallowed by the promise returned from 
  // p.then(). we either have to always register handlers on 
  // the returned promises or we can do the following... 
}, 10);
 
// this time we use .end() which prevents catching thrown errors 
var p = new Promise;
setTimeout(function () {
  p.fulfill(); // throws "shucks" 
}, 10);
return p.then(function(){ throw new Error('shucks') }).end(); // <-- 

Allows direct promise to promise chaining (especially useful by a outside aggregating function). It doesn't use the asynchronous resolve algorithm and so excepts only another Promise as it's argument.

function makeMeAPromise(i) {
  var p = new Promise;
  p.fulfill(i);
  return p;
}
 
var returnPromise = initialPromise = new Promise;
for (i=0; i<10; ++i)
    returnPromise = returnPromise.chain(makeMeAPromise(i));
 
initialPromise.fulfill();
return returnPromise;

###Event names

If you'd like to alter this implementations event names used to signify success and failure you may do so by setting Promise.SUCCESS or Promise.FAILURE respectively.

Promise.SUCCESS = 'complete';
Promise.FAILURE = 'err';

###Luke, use the Source For more ideas read the source, tests, or the mongoose implementation.

MIT