Validation

[ondrejbouda]

No description provided.

[xificurk]

Making the final on message and handler optional opens a whole new can of worms and I'm not sure if that's wise...
Commands and/or handlers can now inherit from each other (which would probably break handler resolver logic). Also, they can be abstract and there is no validation for this.

[ondrejbouda]

In my previous commits I added support for custom validators.

[ondrejbouda]

Making the final on message and handler optional opens a whole new can of worms and I'm not sure if that's wise...
Commands and/or handlers can now inherit from each other (which would probably break handler resolver logic). Also, they can be abstract and there is no validation for this.

This hole is already open. There is nothing that would make you to even use the validators in the first place. I suggest to let this be for now. The default is alright (strict pre-configured validators for commands and events), any other scenario is up to the user.

Maybe we could come up with a solution that somehow ties the validation logic and handler-resolving logic. But for now, I would leave it be.