Kyle Fuller

Specification already states that the source maps contain character maps and not bytes: > This location SHOULD include the characters used to build the parent element > A source map...

After giving this a bit of though, I think we should use byte based source maps from the original source document. Using characters will be problematic for the following reasons:...

Additional note, I think we should provide convinces in Fury/Minim-API-Description to convert a source map to a specific line number as this a common pattern that various editors and tooling...

I think this will happen if the function is invoked without the expected CloudWatch payload as the event. If you're manually testing, an event should look something like the following:...

I was thinking we could introduce a concept of "Server" into API Elements which is similiar to OpenAPI 3. A server could contain a href, and optionally hrefVariables. The href...

Following some discussions in OpenAPI, there is some consensus that `servers` is perhaps a poor name (https://open-api.slack.com/archives/C1137F8HF/p1559929905003800), and `hosts`, `backends` or similiar may be a better fit.

@klokane I think this can be solved with 'One Of', and enum, for example the MSON-pseudo code, the hrefVariables can be represented using selection: + One Of + (object) -...

There's a proof of concept implementation for OpenAPI 2 parser at https://github.com/apiaryio/api-elements.js/commit/50dc4947c1e51a60108f4deef94e3dfd442236da.

I had some discussions with @tjanc who proposed we should re-use the "resource" element because it already contains `href` and `hrefVariables`. Semantically, a server IS a resource. Therefore we can...

> although I think we should make an attempt to unify the codes as well @w-vi Would we still need to provide error codes if the profile/origin link can replace...