chickeneer

I don't use the locale system. But just looking at the code, most everything looks correct. Your report seems ambiguous on how you are creating the problem. Especially the part......

I am fairly confident I know what is happening. `java.lang.reflect.Parameter;` is not able to properly get the type of your method parameter at runtime - related to the generics involved....

Maybe... Some testing needs done given this information. Not 100% certain this gives the full solution. Could get as simple as replacing `this.type = param.getType();` with `this.type = param.getClass();`

Derp. What was I thinking when I typed that. Definitely agree with the linked site not having the needed solution. Theoretically, the would be a plugin compile thing to prevent...

No... I think generics is exactly the problem. It is extending the ISkinCommand interface https://github.com/SkinsRestorer/SkinsRestorerX/blob/c1840af58a68c199e53211390a25236f5aa2cee3/src/main/java/skinsrestorer/shared/interfaces/ISkinCommand.java#L13 Implemented here. https://github.com/SkinsRestorer/SkinsRestorerX/blob/c1840af58a68c199e53211390a25236f5aa2cee3/src/main/java/skinsrestorer/bukkit/commands/SkinCommand.java#L33 From what I can tell. `Parameter#getType` is getting Object from the interface...

It is not a bug. You just want different behavior than what was intended. The original intent when it was written for EMC is as follows. Only our staff could...

My current inclination would be ANOTHER permission to skip over the :confirm. But want to make sure that is what we want...

This sounds sane. We should note that `CommandParameter` is already a class name in the project. This leads me to think that this may not be the most appropriate annotation-name...

Would be nice to also support. `@Subcommand("add name") public void onAdd(CommandSender sender, @CommandParameter("id") NamespacedKey id, String name) { }` as an option to do the same thing as described above.

This is an issue with how the commands are processed by the server. Recently commands are being executed by the brigadier parser which does not allow ~ in the arguments....