based on a customer request, I built something for the flowpack/neos-frontendlogin.
Note: I know that package is intended more as a demo of how to build a login, but it works so well, people actually use it, as it does the job. Similar to the search plugin…
So I built https://github.com/Flowpack/neos-frontendlogin-ui, which offers a module in Neos to create and edit just FE users. It comes with it’s own roles so you can assign permission to use it. It also comes with a CLI command to delete expired users.
Now the questions:
Should I keep this as a separate package? Or integrate it with the Frontend Login package?
Should the CLI command become part of the UserCommandController in Neos?
Or should I hang my head in shame and move it to some other namespace, even?
I personally think it would be good if it would just be part of the old package.
Lately I found splitting packages too much doesn‘t make live easier either. Especially for smaller plugins. And people can simply disable the module anyway.
Of course this makes the package move closer to the functionality of https://github.com/sandstorm/UserManagement/
Do you handle the case already when users are authenticated with LDAP or any other external system? Because then the create/edit/delete makes less sense depending on the setup. But still it would be good to see the data in the backend of the users that logged in at least once.
Well, the sandstorm user management package does not offer a module, so it’s not really overlapping. As far as other login mechanisms go, I didn’t think about that. The UI and CLI only deal with users “specific” to the FE Login package (I am improving that further)…
I am fine either way, I guess the UI is closely bound to the rest anyway as they both need to know about the roles within. I am in general a bit wary about the package and it’s broad assumption of what frontend and backend is, so I am not really in favor of making it even more of a standard than it is already (which would be the result of adding the feature) but I guess that can’t be helped either way. We should just think about how we want to go forward with this decision making of BE / FE, because as it is right now it isn’t feasible long term as we discussed already Karsten.
