Make security pluggable by authentication type
It would be great to refactor the security middleware to be pluggable by authentication type. The goal would be to be able to register a handler, a class or function with an interface we define, to check certain security schemes.
This way people can still use unsuported authentication types (eg. #694), and possibly contribute it back to connexion.
I like the way Starlette handles this in their AuthentictionMiddleware:
https://github.com/encode/starlette/blob/master/starlette/middleware/authentication.py
https://github.com/encode/starlette/blob/master/starlette/authentication.py
https://www.starlette.io/authentication/
I think it's worth investigating whether we can base the connexion security middleware on this.
This would be very helpful, plus 1 from me also. There is also another scenario we would like to have: be able to add on the same application 2 different API Specs with completely different security schemes defined in the spec. For example we could have an end-user API and an administration API which should be able to use completely different authentication logic.
Fixed by #1671