• JS mocha + chai should be used

• Random tests also fail to generate cases when sides < 3

• Description should mention about the correct format of the string

• Example tests should be added

• The equation should be given instead of linking to an external source

JS: Node v12 should be used along with its appropiate assertion tools (Mocha + Chai).
Refer to this and this

Duplicate + The issues below should be unpublished ~~

One of the test cases is broken:

Receiving: [ null, <1 empty item>, undefined, null, 0, '', <1 empty item>, [ <3 empty items> ], [] ]
Expected: [[null],[null,null,0,""],[ [null,null,null],[] ]]

tests are using test.expect instead of chai.assert.deepEqual

No random tests.

No sample tests.

This kata needs its tests completely rewritten to give actually meaningful test results.

Also, no random tests.

JS :

• if this is supposed to be a puzzle, then it lacks the tag for it (and it should not it in the description)

• if it isn't a puzzle, well it should be more clear about how it is to be sorted (see Johan's comment)

Your random test is a joke. Please write proper, strong random tests.

If the original array is dense, for consistency .split() should return an array consisting of a copy of the original array.

How is [-2, 100, 2.0e10, 100.00001] sorted to [100.00001, -2, 2.0e10, 100 ] ?

I just don't understand what sort order this is.