Normative: rename to Iterator.range

[Jack-Works]

API Before (in TypeScript):

var Number: {
    range(start: number, end: number, stepOrOptions: number | NumericRangeOptions<number>): IterableIterator<bigint>
}
var BigInt: {
    range(start: bigint, end: bigint | Infinity, stepOrOptions: bigint | NumericRangeOptions<bigint>): IterableIterator<bigint>
}

API now:

var Iterator: {
    range(start: number, end: number, stepOrOptions: number | NumericRangeOptions<number>): IterableIterator<number>
    range(start: bigint, end: bigint | Infinity, stepOrOptions: bigint | NumericRangeOptions<bigint>): IterableIterator<bigint>
}

[codecov-commenter]

Codecov Report

Merging #59 (b1932a4) into main (aa6d7d6) will increase coverage by 0.60%.
The diff coverage is 100.00%.

@@             Coverage Diff             @@
##             main       #59      +/-   ##
===========================================
+ Coverage   99.39%   100.00%   +0.60%     
===========================================
  Files           1         1              
  Lines         165       145      -20     
  Branches       57        55       -2     
===========================================
- Hits          164       145      -19     
+ Misses          1         0       -1     

Impacted Files	Coverage Δ
polyfill.js	100.00% <100.00%> (+0.60%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[ljharb]

Shouldn’t this coerce start and end to a primitive first, like every other builtin api?

[Jack-Works]

if we coerce it to a primitive, we will lose the possibility to add more overload in the future, for example, supporting a Date range.

// for example:
Iterator.range(new Date('Jan 1 2020'), new Date(), { step: '1 day' })

[ljharb]

That seems like something more appropriate to put on Date (or better, on Temporal) than in this function.

[ljharb]

Another alternative approach is putting Z bar or R bar in a variable, and calling it like a function (also, i don’t think both of these should be Z bar?

[Jack-Works]

what is "Z bar" or "R bar"? 🤔

[ljharb]

ℤ is "Z bar", eg

[ljharb]

Similarly, I’d expect the options to be coerced ToObject when it’s non-nullish, rather than throwing.

[bakkot]

Why?

I guess maybe there's a consistency argument (though I'm not sure with what?) but it seems like that would generally be unintentional.

[Jack-Works]

about coerce option, it's OK to me, but since I decided to not coerce start, I think it will be more consistent.

[ljharb]

Let me step back - what does Temporal do with options bags? We should at least be consistent with that.

[bakkot]

https://tc39.es/proposal-temporal/#sec-getoptionsobject

[ljharb]

alrighty, fair enough. let's inline those same AOs, then, and use them.

[ptomato]

Temporal does what ECMA-402 does (at least, what ECMA-402 does for new APIs; there are some older Intl APIs that coerce primitives to objects). See tc39/ecma402#480

[zloirock]

Iterator.range(new Date('Jan 1 2020'), new Date(), { step: '1 day' })

H̶̢͈̺̲͗͒̈́̈̿͆̋̋͘͝͝ơ̵͓̼̈́̏͑͊̈́͋̈́͒̕͘ļ̶̯̫̭̣̞̙̖̝͎͋̓̄͂͛̐͐̕ŷ̷̨̡̰̘̝͍̻̬͖̰̫͉̣̻̔̀͊͌̃̔̎͛́͛́̑͜͠ ̶̲̫͎̮̍̀́̔s̷̨̨̨̞̘̪̫͇͙͈̫͇̟̣̲̊́̀ͅͅh̸̡̧͕̱͖͙͎̮̐͆̌́̀͌̽̆͛͛̑͛̅̔̕͝į̴̫͈͉̩̺̹̤̹͚̖͇̰̊̊͋͌͑̃̈́̿͆͘̕͝t̸̨̨̛͇̘̠̞̼̏̈́̏̆̀̑̒̔̔͋̽͋̅̀.̴̢̤̖̀̐͂͂̊̂͆͒̇̆͐.̸̼̼̼͍̜͙̗͗̂̓̃̌̍͒̅͛͛̃͐̕̕̕͜͠.̷̡̢͚͙͇̬̳̦̫̫̓́̀̎͜ͅͅ

[zloirock]

It's OK to join number and bigint versions of this method, but consider adding overloads for extra not similar types - it's a kind of god function...

[Jack-Works]

just leave the possibility. not going to do it in the current proposal

[Jack-Works]

I'll merge this first, but this is not the final conclusion of renaming.

I'll ask the committee if they like the rename in the next presentation.