Promises doodle

doodle.js

function retProm() {
 var def = when.defer();

 return def.resolve();
}

console.log('one');
retProm().then(function(){console.log('two'));
console.log('three');

// this will print:
//
// one
// three
// two

yes you are absolutely right.

That's silly and it is not what i am doing.

The (X) that you mention in this gist, represents the fs.readFile operation in my tests.

fs.readFile is an async op. That async op is what i am trying to emulate by adding one more nextTick.

Using your example excluding the nextTick:

var fs = require("fs");

var tickResolvedOn = null;

function readFile(name) {
  var def = when.defer();

  fs.readFile(name, function (err, data) {
    def.resolve(data);
    console.log("(X) an extra tick has passed after resolving");
  });
}

readFile("file.txt").then(function onFulfilled(data) {
    console.log("(Y) data received");
});

console.log("(Z) `then` has returned")

This will print: Z, X, Y

If there was no extraTick in the promise implementation for resolving the promise, it would be: Z, Y, X

No, this could print Z, Y, X. There is no requirement that X come before Y. The spec is still following as long as Y comes after Z.

Let me explain further the difference between our two examples. In yours, when you call def.resolve, then has not yet returned. So the implementation must insert an extra tick before calling onFulfilled, to ensure that then returns first.

In my example, when I call def.resolve, then has already returned, since an async operation has already happened. So the implementation does not have to insert an extra tick before calling onFulfilled, since then has already returned.