-
Notifications
You must be signed in to change notification settings - Fork 57
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Drop @LoginConfig #182
Comments
@rmannibucau Sure, this is more concrete, no doubt about it :-) |
Hey @rmannibucau I've added a 2.0 milestone, though I'll resist as much as possible :-) because we have the users (and likely not only us) using it, unfortunately with more than MP-JWT, so there has to be a strong reason to make them update their applications. Few quick questions:
What is not portable about sticking it to the JAX-RS
Do you mean its optional
MP-JWT ? I certainly know one project where I wonder if a starting point can be to drop the references to all but Cheers |
It is an annotation so if not in the classpath it is ignored, no update to do :).
It is not defined yet ;). It also has the issue to conflict with other EE specs (servlet and security out of my mind) which have their own security mecanism/annotations. Who wins? Who wraps who?
No, the jwt-auth activation itself. With all the proxies we have these days in k8s and friends, it is likely we'll have deployment with JWT Auth and deployments without when going accross cloud providers, using a mp-config toggle is a neat solution to be able to deactivate that feature and can completly replace the annotation.
I guess it is the case of a project going beyond the spec (keycloak usage?) and this would be solved the same way with the mp-config toggle for this app (+ redhat ecosystem using jandex does not need the annotation to exist to see it ;)). |
Hi Romain,
But the legacy project is relying on a user application telling it via
Then its docs can be updated
It can - but we are talking about the new projects here.
No, it is not related to KC. Some immediate clarity would definitely helped though. We can say:
making optional should really work for all |
@sberyozkin before saying I agree or not, can you clarify you propose to make this annotation support optional as "implementation can or not support it" (I'm ok with it) or something else? |
@rmannibucau Yes, optional would imply it, but I'm not 100% to be honest right now. |
@sberyozkin nothing urges but think we should try to relax that soon, if bothersome (= we get negative feedback of making it optional) we could rethink about an useful and not conflicting configuration (can be as simple as enabling to handle a JAXRS interceptors set programmatically in an user feature to enable/disable that, it could even enable to make it dynamic in apps compared to today ;)). |
@rmannibucau Are you OK with renaming this issue to |
@sberyozkin works for me (maybe with another wording since "make @loginConfig optional" sounds like it can be there or not whereas we want to emphasis it can be supported or not). |
@rmannibucau Making it optional is the very first step - once it is optional it effectively means no one has to support it but only those containers which prefer it will do it. By the way, the package.info has:
So there is a reference to But in any case, I'm fine with updating its docs as the 1step as suggested earlier (drop the ref to all but MP JWT methods) |
Yep works for me. Side note: resources are used by jaxrs apps but not always referenced so javadoc does not remove all the ambiguity and clearly encourages the one with EE ecosystem so good to slowly move to dropping it from default required api. |
CC @teddyjtorres. |
Was there a decision reached on this? We're adding MP-JWT for the first time to an app running on WildFly, and I need to know if we should continue to use web.xml or use @LoginConfig and dump our web.xml file. |
Follow up issue of #70 (hoping the scope of this one is clear and narrowed :))
This annotation has a lot of pitfalls:
So overall, there is not a single case which justify this annotation and it does not bring any configuration the implementation would require so there is no blocker to drop it (note that dropping an annotation keeps the application functional without recompilation in terms of API).
Side note: adding a mp-config key to enable/disable this spec can still be done as a consequence of the removal of this annotation IMHO.
The text was updated successfully, but these errors were encountered: