macroexprs on enums

[lenscas]

Due to enums being represented as strings (for good reason.) You can't really attach methods to them.

However, sometimes you want to attach more complex behavior to them. Which right now means you are stuck making free floating functions.

While the ability to attach macroexpressions doesn't take this away entirely, they would allow you to at least make them callable from instances of the enum as if they are methods.

It would also make enums more inline with records and interfaces. As both of those have access to macroexpr based methods.

[Nikita-baghela07]

Good point!, enums as strings keep them simple, but it’s limiting when you want behavior tied to values. Allowing macroexpr methods would make enums more consistent with records and improve readability without breaking existing patterns.

[lenscas]

Iirc there is an issue open regarding the ability to iterate over the variants of an enum.

With macroexprs being able to be attached on enums this could open the door to such a feature.

[Nikita-baghela07]

Attaching macroexprs to enums could naturally enable iteration over variants, making them more consistent with records

[hishamhm]

(Are these, like, LLM replies? They basically repeat @lenscas's points by merely paraphrasing them. The inconsistent punctuation is throwing me off a little, but this user has been posting large non-working PRs full of AI-generated code.)

[hishamhm]

@lenscas could you give an example of the type of code you'd like to write with this feature?

[lenscas]

I ran into it when I was working on macroexpr support for tealr. I currently don't use teal (though I am planning to change that once Tealr is where I want it to be).

It not being on enums caught me off guard as I expected it to be a thing, seeing how it was on both interfaces and records.

However I have had the need to add methods to enums in many projects. And this would be a way to kind of do this.

Say, for user roles you might want a method that checks if the user has at least role X.

While this doesn't fully solve that (as macroexpr has quite a few limits), it would allow bridge the gap at least a bit.