Bug Report: `qs.parse` when parsing a string with empty indexes `[]` in a key.

[leoyli]

I think there is a bug in parseArrays: false option:

normal case with index

qs.parse('a[0]=b', { arrayLimit: 0 }) === { a: [ 'b' ] } // true  <-- expected
qs.parse('a[0]=b', { parseArrays: false }) === { a: { '0': 'b' } } // true  <-- expected

buggy case with empty index

qs.parse('a[]=b', { arrayLimit: 0 }) === '{ a: [ 'b' ] }' // true  <-- expected
qs.parse('a[]=b', { parseArrays: false }) === '{ a: [ 'b' ] }' // true  <-- unexpected

Instead, I will expect if I set parseArrays: false, I will receive no Array:

qs.parse('a[]=b', { parseArrays: false }) === '{ 'a': { '0': 'b' } }' // true

That is because the array parsing is only preserving the order, not by their true indexes in the query, according to the doc:

qs will compact a sparse array to only the existing values preserving their order

Also it is consistent with when other option being set (i.e. treat [] as [0], see below.)

[ljharb]

Regarding bug 1: parseArrays: false should work like this example in the readme:

var noParsingArrays = qs.parse('a[]=b', { parseArrays: false });
assert.deepEqual(noParsingArrays, { a: { '0': 'b' } });

Indeed, this is a bug; I've added it to the tests and will fix it shortly.

Regarding bug 2: can you file that separately, and we can discuss there?

[leoyli]

Filed another bug in #262.