Translation from Conjure-Oxide Model to Minion Model for XYZ Problem

[niklasdewally]

As per #68

[niklasdewally]

@lixitrixi @ozgurakgun What are your thoughts on #[non_exhaustive] on public AST things?
I have started putting some in in here - I think it might be good for backwards compatibility as we add constraints.

https://doc.rust-lang.org/reference/attributes/type_system.html#the-non_exhaustive-attribute

For our purposes, it would effectively always force a _ => Err() branch on any pattern matching involving the ast.

[ozgurakgun]

as far as I understand this forces users to always handle the _ case, is that right?

it could be good for documenting intent, but on the other hand I think we want compiler warnings/errors when we add new cases instead of silently compiling without warnings. code that only cares abotu some of the cases can still use _, but code that is intented to be exhaustive will get a compile time warning/error instead. does that make sense?

[ChrisJefferson]

Note you can use:

        #[allow(unreachable_patterns)]
        _ => panic!(""),

Where you don't want to worry about future cases (and that could be added in commits which add a new thing we temporarily dont want to worry about)

[niklasdewally]

Thank you - I will add these explicitly so that they aren't cargo check --fix 'ed by accident. I think there may still be some potential uses of non_exhaustive in the minion_rs crate - in particular, with error handling and adding new constraints that are not implemented up the stack yet - but that is out of the scope of this PR and would be something to think about when we get back around to polishing up the solver crates.

…

On 16/11/2023 10:55, Christopher Jefferson wrote: Note you can use: |#[allow(unreachable_patterns)] _ => panic!(""), | Where you don't want to worry about future cases (and that could be added in commits which add a new thing we temporarily dont want to worry about) — Reply to this email directly, view it on GitHub <#75 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKQHNE3BSUKZOS674VHEXHLYEXWJVAVCNFSM6AAAAAA7M64UFWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMJUGIYTGNBUGE>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[niklasdewally]

@ozgurakgun I have translation working for XYZ from the flat conjure to minion.

Could do with more testing, but I am relying on private methods and manual construction of all the AST stuff ( i think ::new(), etc are in #32 ?), so these will need to be rewritten anyways.

[ozgurakgun]

minor comments, looks good to me!

[niklasdewally]

@ozgurakgun I can confirm that I can no longer merge PRs to main :)