I've checked URLEncodedFormParser.parse and its behavior appears correct. I havent found the next responsible code in the chain. My suspicion is it makes no sense for the decoder to decode a key as a value maybe. Maybe bogging the key into the value would behave properly by falling through the existing logic. I cant say until I find the chain of execution.
It may even be it doesnt try to decode it as a dictionary of strings as a dictionary with a key and no value is undefined? I'll keep looking.

Decoding string.self does capture it solo or with keyed values but it seems to capture the last string in the URI. I'm still not sure what the solution should be. Perhaps I should check the RFC.

Are you guys sure that "the URI Spec doesn't specify that query parameters must have a value" also mean that it may not have the = sign after the key? In my experience, I often see key= in the wild, but so far I have never seen just key.

Checking the RFC it only indicates that key value pairs are commonly used and does not appear to state any requirements on the structure of the query string other than its termination and how to specify pairs. I think we need to confirm this shold be supported before doing much more work. I think it is a conceptual bug more than an implementation bug perhaps.

In my opinion treating flags as bool arguments should be a workaround for getting flags from query params.

This won't make or break my project, it was more of a curiosity whether this was an oversight/omission in the query parsing code, or an active design decision. I don't have any precedent to refer to in terms of patterns in other projects, so maybe all that suggests this is a bad idea 😅

But the spec is pretty clear about the query component definition:

The query component is indicated by the first question mark ("?") character and terminated by a number sign ("#") character or by the end of the URI.

For reference, I typically defer to MDN docs and/or Javascript implementations as a battle-tested standard, and it appears that JavasScript will parse valueless query parameters.

var u = new URL("https://example.com/?foo=bar&baz")
console.log(u.searchParams) // outputs: URLSearchParams { 'foo' => 'bar', 'baz' => '' }

I get the same result in Safari and NodeJS repl.

Found a workaround via the Vapor.Request.url property. I can get at the raw query that way.

let u: URLComponents? = URLComponents(string: String(describing: request.url))

The Foundation.URLComponents class captures query params without values, so it works exactly as I was hoping.

EDIT: I'll leave it up to the Vapor team to decide if they want to pursue this enhancement or close it.

Cheers ✌

I'm not sure if the current behavior is correct or not. I'm not sure what would change.
Since you have a workaround I vote it can be closed unless more specific action is determined.