what is the difference between thunk, futures, and promises? - promise

There are wiki articles about them: (http://en.wikipedia.org/wiki/Futures_and_promises, http://en.wikipedia.org/wiki/Thunk_(delayed_computation)). But I am not sure what are the exact differences between the three as a programming language concept? Are futures and promises only applicable in concurrent programming?

An example of each, using javascript since everybody can read it.
Please don't use this code in production, use a real library, there are plenty of good ones.
var a, b, c, async_obj; // assume exist
// CommonJS - for reference purposes
try {
async_obj.asyncMethod(a, b, c, function (error1, result1) {
if (error1) {
console.error(error1);
} else {
console.log(result1);
}
});
} catch (ex1) {
console.error(ex1);
}
// Thunk - "a parameterless closure created to prevent the evaluation of an expression until forced at a later time"
function makeThunkForAsyncMethod (cb) {
return function () {
async_obj.asyncMethod(a, b, c, cb);
}
}
var my_thunk = makeThunkForAsyncMethod(function (error1, result1) {
if (error1) {
console.error(error1);
} else {
console.log(result1);
}
});
setTimeout(function () {
try {
my_thunk();
} catch (ex1) {
console.error(ex1);
}
}, 5e3);
// Promise - "a writable, single assignment container which sets the value of the future"
function makePromiseForAsyncMethod () {
var
future_then_cb,
future_catch_cb,
future
;
future = {
then: function (cb) {
future_then_cb = cb;
},
catch: function (cb) {
future_catch_cb = cb;
};
};
try {
async_obj.asyncMethod(a, b, c, function (error1, result1) {
if (error1) {
if (future_catch_cb) {
future_catch_cb(error1)
}
} else {
if (future_then_cb) {
future_then_cb(result1);
}
}
});
} catch (ex1) {
setTimeout(function () {
if (future_catch_cb) {
future_catch_cb(ex1);
}
});
}
return future;
}
// Future - "a read-only placeholder view of a variable"
var my_future = makePromiseForAsyncMethod();
my_future
.then(function (result) {
console.log(result);
})
.catch(function (error) {
console.error(error);
})
;
A Promise chain would be like the above contrived example, but it would work on collections and be more robust.

in functional programming, the most difference between thunk and promise is, thunk is pure while promise is impure.
function thunkDemo() {
return function(callback) {
asyncMethod(someParameter, callback);
};
}
function promiseDemo() {
return new Promise(function(resolve, reject) {
asyncMethod(someParameter, function(err, data) {
if(err) return reject(err);
resolve(data);
});
});
}
when thunkDemo is called, the asyncMethod will not be called until the inner method called, so thunkDemo is pure with no side affect.
when promiseDemo is called, it will call asyncMethod immediatly, which means, it's not pure.

Thunk is a general concept of a small function which is used only to adapt the call or prepare/modify it in some way and then redirect to the proper function. Things like promises, futures, closures, wrappers, stubs or implementations of the concept of virtual function tables in some OO languages (like C++) are just special use cases of thunks (thunks are often used to implement them).

These are quite broad terms, and their usage and interpretations vary with context. So a specific answer can only be given for a specific context.
In javascript promise libraries for example, the terms "deferred" and "promise" are used for what your linked wiki article refers to as "promise" and "future" respectively, in that a "deferred" is the interface for resolving or rejecting the value, and "promise" is the interface for reading it, with some more specifics that allow easily constructing new promises for dependent values, but not actually changing the original promise.
I just stumbled upon the JS library "co" (https://github.com/visionmedia/co).
It was the first time I heard of "thunks", and it seems to be using the term in a slightly different sense from slf's answer, and much more specific than your linked wiki as well. It is not a parameterless function that returns the value, it is a function accepting a callback that will call the callback with the value, often asynchronously.
In this specific library's case, thunks and promises are very close to each other: a promise is an object with a "then" method function that sets a callback for getting the value; a thunk is directly a function that sets a callback for getting the value.

Related

Vuex store action from Promise to async/await

Currently I use promises in the store actions but want to convert it into async/await. This is an example of the store action with promises:
fetchActiveWorkspace (context, workspaceID) {
if (workspaceID) {
return this.$axios.get(`#api-v01/workspaces/workspace/${workspaceID}`)
.then(response => {
context.commit('setActiveWorkspace', response.data)
})
.catch(err => {
throw err
})
} else {
return Promise.resolve(true)
}
},
This fetchActiveWorkspace action is resolved in components because it returns promise. How can I convert this code snippet into a async/await structure and use it in components?
This is how I would try to translate it; take into account that as I have no access to the original code in full context, I cannot try it first-hand to make sure it works; but still, this is how you can use async/await with promises.
// 1. Mark the function as `async` (otherwise you cannot use `await` inside of it)
async fetchActiveWorkspace(context, workspaceID) {
if (workspaceID) {
// 2. Call the promise-returning function with `await` to wait for result before moving on.
// Capture the response in a varible (it used to go as argument for `then`)
let response = await this.$axios.get(`#api-v01/workspaces/workspace/${workspaceID}`);
context.commit('setActiveWorkspace', response.data);
}
// 3. I don't think this is necessary, as actions are not meant to return values and instead should be for asynchronous mutations.
else {
return true;
}
}
You can surround the function's body with try/catch in case you want to capture and handle exceptions. I didn't add it in order to keep things simple and because your promise-based code will just capture and re-throw the exception, without doing anything else.

Implement a "promise always" return function

So i have these 2 functions loadScriptWithBluebird and loadBluebird,
loadScriptWithBluebird is supposed to be generic for the main flow of the application, and will always return a promise. I wonder if it is possible with some kind of a pattern to use loadBluebird inside loadScriptWithBluebird
function loadBluebird(callback){
//load the script and callback ...
script.onload = function() {
callback();
};
}
function loadScriptWithBluebird (src){
if(isBluebirdAvailable()){
return new Bluebird(function(resolve, reject){
//load the script and resolve ...
script.onload = function() {
resolve();
};
});
}else{
//issue starts here since i obviously cannot return inside the callback
loadBluebird(function(){
loadScriptWithBluebird(src)
})
}
},
So the flow of the application will look like this :
loadScriptWithBluebird('jquery').then(function(){
//...
});
Thanks.
No, you can't do that, since bluebird wouldn't yet be available when you need to return the promise. So simply ignore that, and rely on the caller of your loadScript function to already have loaded Bluebird. There's not really a way around this.

Converting callback hell to observable chain

I have been working with a convention where my functions return observables in order to achieve a forced sequential series of function calls that each pass a returned value to their following "callback" function. But After reading and watching tutorials, it seems as though I can do this better with what I think is flatmap. I think I am close with this advice https://stackoverflow.com/a/34701912/2621091 though I am not starting with a promise. Below I have listed and example that I am hoping for help in cleaning up with advice on a nicer approach. I am very grateful for help you could offer:
grandparentFunction().subscribe(grandparentreturnobj => {
... oprate upon grandparentreturnobj ...
});
grandparentFunction() {
let _self = this;
return Observable.create((observer) => {
...
_self.parentFunction().subscribe(parentreturnobj => {
...
_self.childFunction( parentreturnobj ).subscribe(childreturnobj => {
...
observer.next( grandparentreturnobj );
observer.complete();
});
});
});
}
parentFunction() {
let _self = this;
return Observable.create((observer) => {
...
observer.next( parentreturnobj );
observer.complete();
}
}
childFunction() {
let _self = this;
return Observable.create((observer) => {
...
observer.next( childreturnobj );
observer.complete();
}
}
The general rule-of-thumb in RxJS is that you should really try to avoid creating hand-made, custom Observables (i.e., using Observable.create()) unless you know what you're doing, and can't avoid it. There are some tricky semantics that can easily cause subtle problems if you don't have a firm grasp of the RxJS 'contract', so it's usually better to try to use an existing Observable creation function. Better yet, create Observables via applying operators on an existing Observable, and return that.
In terms of specific critiques of your example code, you're right that you should be using .flatMap() to create Observable function chains. The nested Observable.create()s you currently have are not very Rx-like, and suffer from the same problems 'callback hell'-style code has.
Here's an example of doing the same thing your example does, but in a more idiomatic Rx style. doStuff() is our asynchronous function that we want to create. doStuff() needs to call the asynchronous function step1(), chain its result into the asynchronous function step2(), then do some further operations on the result, and return the final result to doStuff()'s caller.
function doStuff(thingToMake) {
return step1(thingToMake)
.flatMap((step1Result) => step2(step1Result))
.map((step2Result) => {
let doStuffResult = `${step2Result}, and then we're done`;
// ...
return doStuffResult;
});
}
function step1(thingToMake) {
let result = `To make a ${thingToMake}, first we do step 1`;
// ...
return Rx.Observable.of(result);
}
function step2(prevSteps) {
let result = `${prevSteps}, then we do step 2`
// ...
return Rx.Observable.of(result);
}
doStuff('chain').subscribe(
(doStuffResult) => console.log(`Here's how you make a chain: ${doStuffResult}`),
(err) => console.error(`Oh no, doStuff failed!`, err),
() => console.debug(`doStuff is done making stuff`)
)
Rx.Observable.of(x) is an example of an existing Observable creator function. It just creates an Observable that returns x, then completes.

Sequelize correctly executing multiple creates + updates

I have a cron job that scrapes a list of items on a website and then inserts or updates records in a database. When I scrape the page, I want to create records for new ones that haven't been created yet, otherwise update any existing ones. Currently I'm doing something like this:
// pretend there is a "Widget" model defined
function createOrUpdateWidget(widgetConfig) {
return Widget.find(widgetConfig.id)
.then(function(widget) {
if (widget === null) {
return Widget.create(widgetConfig);
}
else {
widget.updateAttributes(widgetConfig);
}
});
}
function createOrUpdateWidgets(widgetConfigObjects) {
var promises = [];
widgetConfigObjects.forEach(function(widgetConfig) {
promises.push(createOrUpdateWidget(widgetConfig));
});
return Sequelize.Promise.all(promises);
}
createOrUpdateWidgets([...])
.done(function() {
console.log('Done!');
});
This seems to work fine, but I'm not sure if I'm doing this "correctly" or not. Do all promises that perform DB interactions need to run serially, or is how I have them defined ok? Is there a better way to do this kind of thing?
What you're doing is pretty idiomatic and perfectly fine, the only room for improvement is to utilize the fact Sequelize uses Bluebird for promises so you get .map for free, which lets you convert:
function createOrUpdateWidgets(widgetConfigObjects) {
var promises = [];
widgetConfigObjects.forEach(function(widgetConfig) {
promises.push(createOrUpdateWidget(widgetConfig));
});
return Sequelize.Promise.all(promises);
}
Into:
function createOrUpdateWidgets(widgetConfigObjects) {
return Sequelize.Promise.map(widgetConfig, createOrUpdateWidget)
}
Other than that minor improvement - you're chaining promises correctly and seem to have the correct hang of it.

promises and deferred in ajax and jquery

This is a sample functionality I need, is it possible??. I am facing problem in debugging it
I need those two functions to be ran before I do any other modification how to do it?
function getPromise() {
var deferred = $.Deferred();
$.when(getPromiseother()).done(function() {
deferred.resolve();
});
deferred.getPromiseother()
}
function getPromise() {
var deferrednext = $.Deferred();
$.when($.ajax(somefunction)).done(function() {
deferrednext.resolve();
});
deferrednext.promise()
}
$.when(getPromise).done(function() {
do something
});
This is a very clumsy way to do this - basically a major point of promises is that they compose and chain and you don't need to $.Deferred anywhere that's not converting a callback API to promises.
$.when($.ajax(somefunction), someOtherPromise).then(function(result, other){
// in here both promises have run to fulfillment and data is available
// so place the rest of your code here
});

Resources