Class RequestEngine

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

Throws a FimValidationError for a path with a "." or ".." segment. The resource methods put ids into the path with pathSegment (encodeURIComponent), which leaves those two unchanged, and URL parsing then resolves them: schemas get S1 .. would request /api/v1/schemas/ (the search) and print its result with exit 0. Neither can name a resource. (Percent-encoded forms such as "%2e%2e" are safe: encodeURIComponent turns their "%" into "%25".) An empty segment (// or a trailing /) is rejected the same way; pathSegment already refuses a blank id, so this is a backstop for future path builders.

Throws a FimValidationError for a blank query value, a blank array element, an empty array or a blank parameter name (assertNonBlankParams): the API treats an empty parameter as no filter, so name= would return the unfiltered result. Every search, list and CSV method goes through here, so all are covered.

Perform a GET expecting JSON and parse it into T. With shape, the parsed value must have the documented form (responseShapeProblem), or the call throws a FimParseError: a null, {} or an error envelope answered with a 2xx would otherwise be printed as data, or read as "nothing found".

Perform a GET returning the raw bytes (XML / PDF / CSV downloads). An HTML page answered with a 2xx (a maintenance or login page, a proxy's error page) is a FimParseError, not a download: no endpoint serves HTML, and the CLI would otherwise save it over the user's file with exit 0.

Perform a request with Accept negotiation and transient-error retries.