implement trimLeft and trimRight

[jvoigtlaender]

... since they are not standard, thus not supported in all browsers.

Addresses https://github.com/elm-lang/core/issues/319 and thus issues like elm/package.elm-lang.org#40.

[evancz]

Ah, nice find! Why do you use \s in your regex but mozilla uses a more complex sequence in their polyfill for trim? I know they are trimming both ends, but they are detecting more than just \s.

Also, does trim work?

[evancz]

Ah, nice find! Why do you use \s in your regex but mozilla uses a more complex sequence in their polyfill for trim? I know they are trimming both ends, but they are detecting more than just \s.

Also, does trim work?

[texastoland]

It looks like the EcmaScript specification defines a couple extended Unicode points as whitespace:

WhiteSpace ::
<TAB>
<VT>
<FF>
<SP>
<NBSP>
<ZWNBSP>
<USP>

For practical purposes they're the same.

[texastoland]

It looks like the EcmaScript specification defines a couple extended Unicode points as whitespace:

WhiteSpace ::
<TAB>
<VT>
<FF>
<SP>
<NBSP>
<ZWNBSP>
<USP>

For practical purposes they're the same.

[texastoland]

Also the MDN link reports trim compatibility starting with IE9.

[texastoland]

Also the MDN link reports trim compatibility starting with IE9.

[evancz]

Okay, I guess let's go with this. If folks find it odd that trim and trimLeft work on slightly different kinds of whitespace, let's fix that in another PR.

Thanks @jvoigtlaender and @dnalot!

[evancz]

Okay, I guess let's go with this. If folks find it odd that trim and trimLeft work on slightly different kinds of whitespace, let's fix that in another PR.

Thanks @jvoigtlaender and @dnalot!

[jvoigtlaender]

I had modelled it on both stuff I "found on the internet" as well as your implementation of words:

    function words(str)
    {
        return List.fromArray(str.trim().split(/\s+/g));
    }

Thought it made sense to do trimming by eliminating the same kind of whitespace that is used for finding word boundaries for splitting in that function.

[jvoigtlaender]

I had modelled it on both stuff I "found on the internet" as well as your implementation of words:

    function words(str)
    {
        return List.fromArray(str.trim().split(/\s+/g));
    }

Thought it made sense to do trimming by eliminating the same kind of whitespace that is used for finding word boundaries for splitting in that function.