fix: cleaning the URL properly

[victorforissier]

Doing this pull request because the handling of the access_token on auth breaks user experience. When there are other useful params in the path (e.g. ?docId=xxxxx) the cleanup in the code removes everything instead of just the access_token & refresh_token.

I fixed it by splitting the path and removing just the access_tokens and refresh_tokens and then rebuilding it.

[lowczarc]

Using filter directly on the splitted object instead of allocating an object and deleting its property to then retransform it into an array would save a few lines and avoid 2 object/array allocations.

[lowczarc]

Also, why do you define an anonymous function to call it right after ?

[victorforissier]

1. Yes makes sense. Like that?
2. Wrapped it inside a function because I think it's clearer and delimits that specific action.

[lowczarc]

This is reallocating an anonymous function at runtime every time you're executing it, and it's inconsistent with the rest of the code, pls define your function at global level (and with the function keyword instead of const = () =>...)

[lowczarc]

Also the function name is backward

[victorforissier]

Your memory tests are broken.

[lowczarc]

Your memory tests are broken.

Fixed on main. Don't mind it