SearchParamsParser

SearchParamsParser is the default ObjectParserInterface<URLSearchParams> implementation in the registry and extends ObjectParserAbstract. It converts query strings into a nested plain object with coerced values. It runs in two places: query string parsing (its output is c.search in route handlers) and application/x-www-form-urlencoded request bodies via BodyParser.

Usage

You normally don't call it directly. Given this query string:

?title=hello&count=5&items[0][id]=1&items[1][id]=2&active=true

c.search becomes:

{
	title: "hello",
	count: 5,
	items: [{ id: 1 }, { id: 2 }],
	active: true,
}

Replace it with your own implementation through the Registry.

Key Notation

Keys are split into path segments with bracket/dot notation via parseKey, and the value is written at that path. Purely numeric segments become array indices.

Intermediate containers are created on demand: an array when the next segment is numeric, otherwise a prototype-less object. The result root is prototype-less as well, so hostile keys like __proto__ land as plain own properties.

Value Coercion

Every value goes through tryParseJSON, so values that are valid JSON are coerced:

Coercion follows JSON rules exactly: "007" stays a string (leading zeros are invalid JSON), "1e3" becomes 1000. Remember that this applies to c.search: ?id=5 gives a number, ?id=abc gives a string. Use a schema when the type must be guaranteed.

Repeated Keys

Writing to a slot that already holds a value promotes it to an array; further writes append.

?ids=1           → { ids: 1 }
?ids=1&ids=2     → { ids: [1, 2] }
?ids=1&ids=2&ids=3 → { ids: [1, 2, 3] }

A parameter's type depends on how many times it appears: once gives the value, twice gives an array. Validate with a schema if the shape must be stable.