I am putting together a spellcasting system, and the spells will have visual effects.

I have a Capability which holds the player's active Spell, as well as whichever passive Spells may still be causing residual effects. I would like to access these Spells through a rendering event (I was thinking RenderWorldLastEvent), and perform the relevant rendering.

The problem is that I would like to define these rendering methods in a neat way, rather than dumping them willy-nilly into the event class, especially as there are potentially many Spells, each with potentially different rendering specifications. That would get horrible quite quickly, and I'd have to check each Spell to find the appropriate rendering method. As an alternative, I was thinking of having Spell#render methods which could easily be called from the render event, and would be overridden by extensions of the base class.

These methods, though, would crash a dedicated server when the given Spell was loaded, if I understand things correctly, even though they would never be called.

Looking at the Item class, it seems a solution to a similar problem is accomplished with @OnlyIn, which I think would work here too, but the Documentation says "Only use it if you are overriding a vanilla method that already has @OnlyIn defined. In most other cases ..., use DistExecutor instead."

I've heard some dissuasion around DistExecutor, though, too (and I'm less sure how it works). What would be an appropriate way of separating this code, without producing a mess?

That's a pretty nice way of doing it, thanks!