Class RequestEngine

Build a fully-qualified URL from a path and optional query parameters.

GET a feed path and parse the XML reply. The Bundesrat feeds require the view=renderXml render parameter; without it the server returns its HTML shell, which we detect and reject with a helpful BundesratParseError.

NOTE: the response Content-Type does not decide whether the body is a feed. The Government Site Builder CMS is inconsistent about the media type (feeds have been seen as text/plain, application/xml, text/html), so we sniff the body — an <!doctype html> / <html> prefix is the HTML-shell guard — rather than trust the header. Don't "harden" this by validating Content-Type; it would reject valid feeds. Only its charset is used, to decode a body that declares no encoding (decodeXml).

Like getXml, but also returns the root element's name.

Perform a GET with Accept negotiation and transient-error retries. Redirects are NOT followed — a 3xx surfaces as an error (the canonical host answers directly).