qs.parse omitting empty string keys

[Coobaha]

Hi, it will be nice to have this part configurable:

https://github.com/ljharb/qs/blob/542a5c7ff88d7229efa2e22c7c8a7d69375f5e72/lib/parse.js#L150-L152

current output:

qs.parse("=1&=2") === {}
qs.parse(" =1& =2") === { " ": ["1","2"] }

expected output :

qs.parse("=1&=2", { allowEmptyKeys: true }) === { "": ["1","2"] }

If you are ok - I can open a PR

[ljharb]

Please link me if I'm wrong, but I don't believe the spec permits empty keys in a query string. What's your use case/software where that's expected?

[Coobaha]

Thanks for the fast reply, I was using this as a reference:

new URLSearchParams('=1').getAll('') // ['1']

Our use case is that we need to evaluate user input in HTTP client editor, in some cases key might be not present yet or really evaluates to empty string. As a workaround I am assigning empty key names to temporal key and after parsing replacing them back :D This also requires preserving initial keys order, so hack becomes quite complex..

I can also imagine that some server/api might use this empty param name, even if it is not part of any RFC

[ljharb]

"USP supports it" is a good argument, although your use cases are not convincing - if the key's not present yet, then the value isn't either, and I'm not clear on why the key would really evaluate to an empty string.

The PR adding this would have to add it to both parse and stringify, so that there could be a round trip.

Additionally, what happens with =1&=2, in the arrayFormat repeat vs brackets? What about []=1&[]=2?

[Coobaha]

Great! i will open a pr then soon and we can continue there 🙏

[ljharb]

To be clear: a PR will need:

• to add an option to both parse and stringify
• to document in the readme
• to have tests, for both parse and stringify, for at least:
  • repeat keys with all array formats
  • bracketed keys with all array formats
  • a lone = (and =& and &= and &=&)

[Coobaha]

@ljharb opened #433 👍