intersperse for highland integration

[buzzdecafe]

No description provided.

[megawac]

Use push instead of concat

[davidchambers]

Didn't we decide to use xs[xs.length] = x? See #891.

I'd be happy for us to reverse that decision, but were we to do so it would be orthogonal to this change.

[megawac]

Push is just better in this case David
On Apr 21, 2015 9:51 PM, "David Chambers" notifications@github.com wrote:

In src/intersperse.js
#1043 (comment):

• * @param {*} separator The element to add to the list.
• * @param {Array} list The list to be interposed.
• * @return {Array} The new list.
• * @example
• * R.interpserse('n', ['ba', 'a', 'a']); //=> ['ba', 'n', 'a', 'n', 'a']
• */
  +module.exports = _curry2(_checkForMethod('intersperse', function intersperse(separator, list) {
• var out = [];
• var idx = -1;
• var length = list.length;
• while (++idx < length) {
• if (idx === length - 1) {

•  out = out.concat(list[idx]);
  

• } else {

•  out = out.concat(list[idx], separator);
  

Didn't we decide to use xs[xs.length] = x? See #891
#891.

I'd be happy for us to reverse that decision, but were we to do so it
would be orthogonal to this change.

—
Reply to this email directly or view it on GitHub
https://github.com/ramda/ramda/pull/1043/files#r28838931.

[davidchambers]

Push is just better in this case

Because we're pushing two values? Fair enough.

I'd like to use push in all cases, for the record. :)

[CrossEye]

So would I. I'm really not sure how we ended up in our current state.

[CrossEye]

Agree with @megawac on push/concat, but other than that, LGTM.

[davidchambers]

Given that the return value has type [a] the separator must have type a.

[buzzdecafe]

hmmm. really, the input array could be [*] and the return could be [*], so the whole sig probably should be * -> [*] -> [*].

[CrossEye]

Do we describe these with the most generic signature that fits or the expected one? Clearly a -> [a] -> [a] is nicer and captures the likely use of the function. But there is nothing to prevent, say,

intersperse (new Date, ['a', 1, /\d+/, {}, []]):

I don't know that we have any choice but to document the most generic version, but it certainly doesn't make me happy.

[davidchambers]

I prefer a -> [a] -> [a] here. I prefer to capture the intended use of the function. If we're completely honest we can end up with types such as * -> * -> * which aren't particularly helpful.

[buzzdecafe]

I prefer a -> [a] -> [a] here.

i'm fine with that too. will revise when able

[CrossEye]

Go ahead, twist my arm.

Ok, ok, I'm convinced!

[mpajunen]

Typo, intersperse

[buzzdecafe]

nice catch, thanks

[buzzdecafe]

ping

👶 ➡️ 👴

[davidchambers]

🌳

[anton-rudeshko]

Great and long awaited addition, thanks! I implemented same extension in lodash, but it wasn't accepted.