Function.prototype.partial

Isiah Meadows isiahmeadows at gmail.com
Mon Oct 5 10:21:56 UTC 2015


I agree that it would be nice. I frequently come up with a helper
function for that, although my lazy fingers don't feel like typing 7
letters each time to create a partially applied function...

```js
// Maybe Function.prototype.part?
list.map(create.part('type'))
```

On Sun, Oct 4, 2015 at 4:07 PM, Andrea Giammarchi
<andrea.giammarchi at gmail.com> wrote:
> for code sake, the dynamic should copy and push, not just push
>
> ```js
> Object.defineProperty(
>   Function.prototype,
>   'partial',
>   {
>     writable: true,
>     configurable: true,
>     value: function partial() {
>       for (var f = this, a = [], i = 0; i < arguments.length; a[i] =
> arguments[i++]);
>       return function () {
>         var args = a.slice(0);
>         args.push.apply(a, arguments);
>         return f.apply(this, a);
>       };
>     }
>   }
> );
> ```
>
> Regards
>
>
> On Sun, Oct 4, 2015 at 10:04 PM, Andrea Giammarchi
> <andrea.giammarchi at gmail.com> wrote:
>>
>> I think his point was that in the functional programming way he is
>> thinking about, you don't care about `this` at all so you can shim `partial`
>> like:
>>
>> ```js
>> Object.defineProperty(
>>   Function.prototype,
>>   'partial',
>>   {
>>     writable: true,
>>     configurable: true,
>>     value: function partial() {
>>       for (var a = [null], i = 0; i < arguments.length; a[++i] =
>> arguments[i - 1]);
>>       return this.bind.apply(this, a);
>>     }
>>   }
>> );
>> ```
>>
>> The dynamic `this` alternative seems quite straight forward anyway
>>
>> ```js
>> Object.defineProperty(
>>   Function.prototype,
>>   'partial',
>>   {
>>     writable: true,
>>     configurable: true,
>>     value: function partial() {
>>       for (var f = this, a = [], i = 0; i < arguments.length; a[i] =
>> arguments[i++]);
>>       return function () { a.push.apply(a, arguments); return
>> f.apply(this, a); };
>>     }
>>   }
>> );
>> ```
>>
>> Regards
>>
>>
>>
>>
>> On Sun, Oct 4, 2015 at 8:04 PM, Jordan Harband <ljharb at gmail.com> wrote:
>>>
>>> (note, it couldn't be a direct proxy to `bind` because `bind(null)`
>>> permanently sets the "this" value (ie, `(function () { return this
>>> }).bind(null).call([])` will return null or the global object, not the empty
>>> array. the polyfill would be a bit trickier, and couldn't use "bind" under
>>> the hood without changes to bind itself)
>>>
>>> On Sun, Oct 4, 2015 at 8:06 AM, Tim Ruffles <oi at truffles.me.uk> wrote:
>>>>
>>>> It'd be nice to add Function.prototype.partial, which is exactly like
>>>> `.bind` without the `this`-setting first argument.
>>>>
>>>> Adding it would have benefits:
>>>>
>>>> - `.bind(null, ...` is newbie confusing and ugly in functional code[1]
>>>> - `.bind(null, ...` is a very common idiom[2]
>>>> - `.partial` is intention revealing: it's clear you're writing
>>>> functional code that doesn't care about `this`
>>>>
>>>> Adding it seems low cost, and is backwards compatible:
>>>>
>>>> - implementors could literally make it a proxy to `.bind` if they want
>>>> to get it done fast
>>>> - shimmable
>>>> - not new functionality, just a subset of existing functionality
>>>>
>>>> Cheers,
>>>>
>>>> Tim
>>>>
>>>> @timruffles
>>>>
>>>> - [1] I've been teaching newbies Javascript/Node.js for years, and if
>>>> you take a 'functions first' approach it's complexity you have to tell them
>>>> to ignore
>>>> - [2] see github -
>>>> https://github.com/search?l=javascript&q=%22.bind%28null%22&ref=searchresults&type=Code&utf8=%E2%9C%93
>>>>
>>>> _______________________________________________
>>>> es-discuss mailing list
>>>> es-discuss at mozilla.org
>>>> https://mail.mozilla.org/listinfo/es-discuss
>>>>
>>>
>>>
>>> _______________________________________________
>>> es-discuss mailing list
>>> es-discuss at mozilla.org
>>> https://mail.mozilla.org/listinfo/es-discuss
>>>
>>
>
>
> _______________________________________________
> es-discuss mailing list
> es-discuss at mozilla.org
> https://mail.mozilla.org/listinfo/es-discuss
>



-- 
Isiah Meadows


More information about the es-discuss mailing list