Curry: Add functions curry2 through curry5 #289

laduke · 2016-11-19T06:47:51Z

Hi, first pull request here. Thanks for the hand-holding. This closes #277.
Things not complete or I'm not sure about:

Examples for curry3 - 5.
Looks like all of the examples in the readme are one liners.
Documentation text is probably too short, but I'm not the one that should be explaining currying to anyone.
I probably skipped some test cases. Is there an S.__ or similar that needs to be tested?
The definition forvar e = $.TypeVariable('e'); is skipped and I don't know why. So curry4 and 5 don't have an 'e': S.curry4 = def('curry4', {}, [$.Function, a, b, c, d, f], curry4);
Not sure what variable names people traditionally use for functions of large arity.
I have function(v, w, x, y, z)...
Should I make the README and commit, or does that happen later?

davidchambers · 2016-11-19T11:16:40Z

Thanks for the pull request, @laduke! It's exciting to have received pull requests from three new code contributors in the past week!

I edited this pull request's description to include the substring closes #277 so the issue this pull request addresses will automatically be closed when this pull request is merged. See Closing Issues via Pull Requests for details. :)

davidchambers · 2016-11-19T11:21:57Z

The definition forvar e = $.TypeVariable('e'); is skipped and I don't know why.

We simply haven't needed this so many unary type variables before, but we're free to define e. :)

Should I make the README and commit, or does that happen later?

The readme gets updated automatically at release time. See sanctuary-js/sanctuary-set#1 (comment) if you're interested in the details.

davidchambers · 2016-11-19T11:25:46Z

index.js

+  //# curry2 :: ((a, b) -> c) -> a -> b -> c
+  //.
+  //. Curry a binary function.
+  //. ```javascript


Please include an "empty" (//.) line between the description and the code block.

davidchambers · 2016-11-19T12:05:22Z

index.js

+  //. ```javascript
+  //. > todo
+  //. todo
+  //. ```


Here's an example which does not seem contrived:

> global.toUrl = S.curry5((protocol, creds, hostname, port, pathname) => . protocol + '//' + . S.maybe('', _ => _.username + ':' + _.password + '@', creds) + . hostname + . S.maybe('', S.concat(':'), port) + . pathname . ) > toUrl('https:')(S.Nothing)('example.com')(S.Just('443'))('/foo/bar') 'https://example.com:443/foo/bar' > toUrl('https:', S.Nothing, 'example.com', S.Just('443'), '/foo/bar') 'https://example.com:443/foo/bar'

davidchambers · 2016-11-19T12:08:20Z

test/curry2.js

+  eq(typeof S.curry2, 'function');
+  eq(S.curry2.length, 3);
+
+  var curried = S.curry2(source);


Let's define the function here:

var curried = S.curry2(function(x, y) { return x * y; });

davidchambers · 2016-11-19T12:12:49Z

index.js

@@ -632,6 +632,54 @@

  //. ### Function

+  //# curry2 :: ((a, b) -> c) -> a -> b -> c
+  //.
+  //. Curry a binary function.


I like these short descriptions. For consistency we should use Curries a binary function. Also, we could be more specific about which function gets curried:

Curries the given binary function.

laduke · 2016-11-19T20:33:39Z

Thank you for your patience @davidchambers.

davidchambers · 2016-11-19T22:39:09Z

index.js

+  //. > global.createRect = S.curry4((x, y, width, height) => (
+  //. .   {x, y, width, height}
+  //. .   )
+  //. . )


I'd prefer this formatting:

> global.createRect = S.curry4((x, y, width, height) => . ({x, y, width, height}) . )

davidchambers · 2016-11-19T22:40:57Z

index.js

+  //.
+  //. ```javascript
+  //. > R.map(S.curry2(Math.pow, 10), [1, 2, 3])
+  //. [ 10, 100, 1000 ]


Let's use [10, 100, 1000] here (without the extra spaces).

davidchambers · 2016-11-19T22:41:36Z

index.js

+  //. createRect
+  //.
+  //. > createRect(0)(0)(10)(10)
+  //. { x: 0, y: 0, width: 10, height: 10 }


Let's use {x: 0, y: 0, width: 10, height: 10} here (without the extra spaces).

laduke · 2016-11-19T23:48:19Z

I've become spoiled by jsfmt. Sorry about those.

davidchambers · 2016-11-20T00:04:10Z

test/curry5.js

+  eq(curried(1, 2, 3, 4, 5), 15);
+  eq(curried(1)(2)(3)(4)(5), 15);
+  eq(curried(1, 2)(3, 4)(5), 15);
+  eq(curried(1)(2, 3)(4, 5), 15);


We might as well enumerate all the combinations. I wrote a script to make this easier:

'use strict'; const pow2 = exp => Math.pow(2, exp); const generateAssertions = (xs, f) => { const l = xs.length; return Array(pow2(l - 1)).join('x').split('x').reduce((s, _, m) => xs.reduce((s, x, n) => `${s}${x}${n === l - 1 ? '' : m % pow2(l - n - 1) >= pow2(l - n - 2) ? ')(' : ', '}`, `${s} eq(curried(` ) + `), ${f(...xs)});\n`, '' ); }; process.stdout.write(generateAssertions([1, 2, 3, 4, 5], (v, w, x, y, z) => v + w + x + y + z));

$ node generate-assertions.js eq(curried(1, 2, 3, 4, 5), 15); eq(curried(1, 2, 3, 4)(5), 15); eq(curried(1, 2, 3)(4, 5), 15); eq(curried(1, 2, 3)(4)(5), 15); eq(curried(1, 2)(3, 4, 5), 15); eq(curried(1, 2)(3, 4)(5), 15); eq(curried(1, 2)(3)(4, 5), 15); eq(curried(1, 2)(3)(4)(5), 15); eq(curried(1)(2, 3, 4, 5), 15); eq(curried(1)(2, 3, 4)(5), 15); eq(curried(1)(2, 3)(4, 5), 15); eq(curried(1)(2, 3)(4)(5), 15); eq(curried(1)(2)(3, 4, 5), 15); eq(curried(1)(2)(3, 4)(5), 15); eq(curried(1)(2)(3)(4, 5), 15); eq(curried(1)(2)(3)(4)(5), 15);

You could run this for each of the other 🍛 functions as well.

davidchambers · 2016-11-20T00:08:33Z

We should make it clear that the functions are curried Ramda-style. I'd like to see two applications of each function. For example:

> toUrl('https:')(S.Nothing)('example.com')(S.Just('443'))('/foo/bar')
'https://example.com:443/foo/bar'

> toUrl('https:', S.Nothing, 'example.com', S.Just('443'), '/foo/bar')
'https://example.com:443/foo/bar'

davidchambers · 2016-11-20T00:15:31Z

index.js

+  function curry5(f, v, w, x, y, z) {
+    return f(v, w, x, y, z);
+  }
+  S.curry5 = def('curry5', {}, [$.Function, a, b, c, d, e, f], curry5);


I'd like to use r (result?) in place of f, because in another pull request f will become a unary type variable. In other words, its type will be Type -> Type rather than Type, and it will no longer be suitable in this context. Using a different type variable now will save me from having to change it in the other pull request.

laduke · 2016-11-20T02:08:32Z

Chippin' away. Thanks for that script.
I now see you wrote that curry5 example with 2 different ways of applying.

davidchambers · 2016-11-21T01:43:08Z

test/curry3.js

+  eq(curried(1, 2, 3), 6);
+  eq(curried(1)(2)(3), 6);
+  eq(curried(1, 2)(3), 6);
+  eq(curried(1)(2, 3), 6);


For consistency, let's use the script to generate the assertions for curry2 and curry3 as well. You have all the combinations, but my obsessive streak wants to see them consistently ordered. 😜

I also changed them to use + instead of * in the curried function. ;)

Works for me. 👍

davidchambers · 2016-11-21T01:46:29Z

Looking good, @laduke! After addressing the final comment could you squash the commits and rebase on master? We'll then be ready to merge, I think. :)

davidchambers · 2016-11-21T02:40:51Z

test/curry5.js

+  eq(curried(1)(2)(3, 4)(5), 15);
+  eq(curried(1)(2)(3)(4, 5), 15);
+  eq(curried(1)(2)(3)(4)(5), 15);
+});


Could we have an empty line after the last assertion? Sorry for the pedantry. No more requests, I promise!

Please use git commit --amend when making the change. :)

I deserve every word of it.

davidchambers · 2016-11-21T02:55:23Z

🌳 Terrific first contribution! Thanks for all your work on this. :)

laduke · 2016-11-21T02:55:49Z

🎉

davidchambers · 2017-02-13T00:49:38Z

$ npm install [email protected]

$ node
> const S = require('sanctuary')
undefined
> S.curry2
curry2 :: ((a, b) -> c) -> a -> b -> c
> S.curry3
curry3 :: ((a, b, c) -> d) -> a -> b -> c -> d
> S.curry4
curry4 :: ((a, b, c, d) -> e) -> a -> b -> c -> d -> e
> S.curry5
curry5 :: ((a, b, c, d, e) -> r) -> a -> b -> c -> d -> e -> r

davidchambers reviewed Nov 19, 2016

View reviewed changes

davidchambers reviewed Nov 20, 2016

View reviewed changes

davidchambers reviewed Nov 21, 2016

View reviewed changes

curry2: Add functions curry2, curry3, curry4, curry5.

ea8b184

davidchambers merged commit 4421ef3 into sanctuary-js:master Nov 21, 2016

davidchambers mentioned this pull request Nov 25, 2016

logic: make S.and and S.or monomorphic, and remove S.xor #279

Merged

davidchambers mentioned this pull request Feb 13, 2017

v0.12.0 release notes #332

Closed

davidchambers mentioned this pull request Jun 20, 2017

remove functions which take uncurried functions #416

Closed

davidchambers mentioned this pull request Aug 10, 2020

function: remove S.curry2, S.curry3, S.curry4, and S.curry5 #692

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Curry: Add functions curry2 through curry5 #289

Curry: Add functions curry2 through curry5 #289

laduke commented Nov 19, 2016 •

edited

Loading

davidchambers commented Nov 19, 2016

davidchambers commented Nov 19, 2016

davidchambers Nov 19, 2016

davidchambers Nov 19, 2016

davidchambers Nov 19, 2016

davidchambers Nov 19, 2016

laduke commented Nov 19, 2016

davidchambers Nov 19, 2016

davidchambers Nov 19, 2016

davidchambers Nov 19, 2016

laduke commented Nov 19, 2016

davidchambers Nov 20, 2016

davidchambers commented Nov 20, 2016

davidchambers Nov 20, 2016

laduke commented Nov 20, 2016 •

edited

Loading

davidchambers Nov 21, 2016

laduke Nov 21, 2016

davidchambers Nov 21, 2016

davidchambers commented Nov 21, 2016

davidchambers Nov 21, 2016

laduke Nov 21, 2016

davidchambers commented Nov 21, 2016

laduke commented Nov 21, 2016

davidchambers commented Feb 13, 2017

Curry: Add functions curry2 through curry5 #289

Curry: Add functions curry2 through curry5 #289

Conversation

laduke commented Nov 19, 2016 • edited Loading

davidchambers commented Nov 19, 2016

davidchambers commented Nov 19, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

laduke commented Nov 19, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

laduke commented Nov 19, 2016

Choose a reason for hiding this comment

davidchambers commented Nov 20, 2016

Choose a reason for hiding this comment

laduke commented Nov 20, 2016 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

davidchambers commented Nov 21, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

davidchambers commented Nov 21, 2016

laduke commented Nov 21, 2016

davidchambers commented Feb 13, 2017

laduke commented Nov 19, 2016 •

edited

Loading

laduke commented Nov 20, 2016 •

edited

Loading