Improve performance of query filters (especially on large bodies)

[aukevanleeuwen]

Fixes: #1838

Rewrite two QueryFilters with an implementation based on StringTokenizer rather than a possible O(n^2) regex (for more information see the linked issue).

Description

I've looked at the description for a application/x-www-form-urlencoded encoded body here: https://url.spec.whatwg.org/#urlencoded-parsing and tried to follow those lines. I did not put in the percentage encoding, since that wasn't in the current implementation and I didn't want to break backwards compatibility.

I've added a few test cases that weren't covered (and also not all handled correctly in the previous implementation), but I think they are rather obscure. Where normally you have key=value, it's also possible to just have key, if there is no = there, key is treated as the key and there is no value. Also key= and =value are allowed even, resulting in an empty key or empty value respectively.

Motivation and Context

Fixes #1838

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have added tests to cover my changes.

[aukevanleeuwen]

@whiskeysierra perhaps? :-)

[2k16daniel]

Hoping that this commit will be merged

[kasmarian]

Thank you for the PR @aukevanleeuwen, it looks great, with StringTokenizer the performance will be way faster!

[kasmarian]

In order for the PR to be compliant with our merging rules, could you please make sure your commits are signed?

[aukevanleeuwen]

@kasmarian ✅

[kasmarian]

👍

[lukasniemeier-zalando]

👍