Classifies a request content-type as one of the two form encodings.

The media type is matched EXACTLY against the two supported types after case-folding, and parameters are parsed separately through the shared parseContentType — which is also what parseMultipart reads its boundary from, so the classifier and the parser cannot disagree about a header. A substring search over the raw value accepted three shapes that are not forms (each measured): a suffixed media type (application/x-www-form-urlencoded-v2), a supported type appearing inside an unrelated QUOTED parameter (text/plain; note="…urlencoded"), and boundary= matching inside a different parameter NAME (xboundary=q).

A multipart/form-data type carrying no boundary parameter — or one whose value is empty — is undefined: the body is not parseable as a form, and reporting an encoding that cannot be parsed would hand the caller a guaranteed throw.

Pure — this is the one classifier the upload middleware's multipart guard and the CSRF verifier's form guard both read, replacing their private includes() copies.