URLSearchParams

Requirements

• Export class SearchParams as named and default. Accept a query string (default empty), removing at most one leading question mark. This is a limited URLSearchParams-style polyfill, not a full URL parser.
• Split on ampersands, ignore empty segments, and split each entry on its first equals sign. A missing equals sign means an empty value. Preserve duplicate keys, empty keys, and entry order.
• Decode plus signs as spaces before percent decoding. Percent-encoded plus signs stay literal plus signs. Valid UTF-8 encoded strings are in scope; malformed percent encoding must throw URIError. This error behavior intentionally differs from native URLSearchParams.
• Implement get(name) returning the first value or null, getAll(name) returning a fresh array, has(name), and read-only size counting entries including duplicates.
• Implement entries(), keys(), values(), and default iteration in entry order. entries and default iteration yield fresh [key, value] pairs. forEach(callback, thisArg) calls callback(value, key, instance) with the specified receiver.
• Implement toString() without a leading question mark: encode each key/value using UTF-8 form encoding, represent spaces as plus signs, keep ASCII letters/digits/*/hyphen/dot/underscore unescaped, and percent-encode other characters with uppercase hex.
• Do not use native URLSearchParams or URL. Mutation APIs, object/iterable constructors, live iteration under mutation, and full malformed-input recovery are outside scope.

Examples

Input: const params = new SearchParams("?tag=react&tag=vue&q=hello+world");
params.get("tag");
params.getAll("tag");
Output: "react" and ["react", "vue"].

Input: params.get("q");
params.toString();
Output: "hello world" and "tag=react&tag=vue&q=hello+world".