Why :malformed? check before :known-content-type?

[vspinu]

I am going through liberator put-all-together example and I am puzzled as of why json parsing occurs so early in the decision tree (:malformed? node). The content header is checked in the later stage (:known-content-type?). So, if content is not json, I am parsing it for nothing.

It looks to me that either malfromed? should occur later in the decision tree, or the tutorial's json parsing should occur in processable? node.

[ordnungswidrig]

You've found an issue. As stated in the documentation :malformed? is for HTTP protocol level errors when parsing the request. One can argue whether invalid data in the request entity ("body") counts as an malformed request or not. In general I would advise to use :processable? for this but unfortunately the status code 422 "not-processable" is not part of the HTTP spec but defined in a WebDAV extension.

IIRC the tutorial was written at a time when :processable? did not yet exist in liberator. In any case the :malformed? function must take the request content-type into account before trying to parse it as JSON.