How to return a different data type in same UserData's methods

[blame2020]

In typically, '==' returns a bool. and '+' returns same type as l/r value.
But it looks like no such implementation is possible because add_methods can only return the same type.
Any ideas to solve this problem?

impl rlua::UserData for XXX {
    fn add_methods<'lua, M: rlua::UserDataMethods<'lua, Self>>(methods: &mut M) {

        methods.add_meta_method(rlua::MetaMethod::ToString, |_, this, _: ()| {
            Ok(Self(...))
        });

        methods.add_meta_method::<'lua, bool>(rlua::MetaMethod::Eq, |_, this, value: Decimal| {
            Ok(bool) // NG
        });
    }
}

[jugglerchris]

Hi,
Can you be more specific about what you're trying to do and how it is going wrong?
For example this works:

use rlua;

#[derive(Clone)]
struct MyType(String);

impl rlua::UserData for MyType {
    fn add_methods<'lua, M: rlua::UserDataMethods<'lua, Self>>(methods: &mut M) {
        methods.add_meta_method(rlua::MetaMethod::Add, |_, this, other: MyType| {
            Ok("some string")
        });
        methods.add_meta_method(rlua::MetaMethod::ToString, |_, this, _: ()| {
            Ok(this.0.clone())
        });
    }
}

fn main() -> rlua::Result<()> {
    let lua = rlua::Lua::new();

    lua.context(|lua_ctx| {
        let globals = lua_ctx.globals();
        globals.set("x", MyType("xx".into()))?;
        globals.set("y", MyType("yy".into()))?;
        lua_ctx.load(r#"
            print(x)
            print(y)
            print(x+y)
            "#).exec()?;
        Ok(())
    })?;
    Ok(())
}

The Add metamethod above returns a plain string instead of the type of the userdata argument.

Note that there are some restrictions - for example the result of the comparison operators (including ==) is always converted to bool, but that's a feature of Lua itself and not imposed by rlua.

[blame2020]

I'm sorry.
It was a misunderstanding.
Thanks to you, I was able to organize it.
The "Ok (bool)" part was actually written as "Ok (Self (true))".
I was able to do what I wanted to do.

[jugglerchris]

Nothing to be sorry about! And glad you've solved your problem.