fix: keep orignal order if inputs are nil for caseInsensitiveCmp

[richardo2016]

in function caseInsensitiveCmp, to keep original order when other items with specified property undefined, we should use logic like below:

function caseInsensitiveCmp (a: any, b: any) {
    if (!a && !b) return 1
    else if (!a) return 1
    else if (!b) return -1
    ...
}

simplify it, we got

function caseInsensitiveCmp (a: any, b: any) {
    if (!a) return 1
    if (!b) return -1
    ...
}

[harttle]

Your newly added case failed, see travis log. I tried shopify/liquid but find it rather confusing:

// scope in ruby
scope = { "students" => [
    { "name" => 'bob' },
    { "name" => 'alice', "age" => 2 },
    { "name" => 'amber' }
]}

// template
{{ students | sort_natural: "age" | map: "name" | join }}

// output
alice amber bob

It seems that those items without the specified property are moved to the tail, and they are re-ordered alphabetically. I'm not sure whether keys in Ruby Map are ordered or not, but certainly it's not behaving the same way as JavaScript.

[harttle]

🎉 This issue has been resolved in version 9.11.0 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀

[harttle]

Finally I find out that in Ruby all items with a null property value goes to the tail of the sorted list. I altered the test case a bit and return 0 for a == null && b == null to keep the original order. There're still differences in sorting and property iteration order, I added some notes in README.md.

Though this PR can't be merged due to the test case not passing, thank you for pointing this out and initiating the test case, @all-contributors please add richardo2016 for code

[allcontributors]

@harttle

I've put up a pull request to add @richardo2016! 🎉

[richardo2016]

Finally I find out that in Ruby all items with a null property value goes to the tail of the sorted list. I altered the test case a bit and return 0 for a == null && b == null to keep the original order. There're still differences in sorting and property iteration order, I added some notes in README.md.

Though this PR can't be merged due to the test case not passing, thank you for pointing this out and initiating the test case, @all-contributors please add richardo2016 for code

thx for your attention to this feature!