What would be the desired behavior of accessing attributes/methods in non-secure context? Shouldn't the context object itself be exposed (which the attributes/methods belong to)? Or Should it be exposed and the actual access/invocation to them throw?
Currently service worker spec expose the object and throw when the attributes/methods are accessed.
This is being hammered out in the [SecureContext] IDL attribute. It won't be defined in this document except by reference: https://github.com/heycam/webidl/pull/65
What would be the desired behavior of accessing attributes/methods in non-secure context? Shouldn't the context object itself be exposed (which the attributes/methods belong to)? Or Should it be exposed and the actual access/invocation to them throw?
Currently service worker spec expose the object and throw when the attributes/methods are accessed.
Related issue in IDL: https://github.com/heycam/webidl/pull/65 Background: