Intrinsic-test: Updated the Constraint enum to support discrete values

[madhav-madhusoodanan]

Context

This PR is part of the changes that were originally made in the x86 extension PR #1814 for intrinsic-test, and is intended to unblock efforts to support other architectures too.

Further context: #1884 (comment)

cc: @folkertdev @Amanieu

[rustbot]

r? @folkertdev

rustbot has assigned @folkertdev.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[madhav-madhusoodanan]

@folkertdev I remember our discussion on whether there might be a better name than "Constraint".

Essentially, this feature is present for arguments that have a restricted set of legal values.

Would a word like "Restriction" or so be more appropriate?

[folkertdev]

I was thinking something more like ConstArgument? That is slightly rust-specific, but really these are just arguments, they just need to be known at compile-time.

Weirdly CI fails in loongarch, that's strange.

[madhav-madhusoodanan]

I was thinking something more like ConstArgument? That is slightly rust-specific, but really these are just arguments, they just need to be known at compile-time.

In this context though, would it be appropriate?

/// An argument for the intrinsic.
#[derive(Debug, PartialEq, Clone)]
pub struct Argument<T: IntrinsicTypeDefinition> {
    /// The argument's index in the intrinsic function call.
    pub pos: usize,
    /// The argument name.
    pub name: String,
    /// The type of the argument.
    pub ty: T,
    /// Any constraints that are on this argument
    pub constraint: Option<Constraint>,
}

Weirdly CI fails in loongarch, that's strange.

Let me try making another commit, perhaps this might be a one-off occurrence with the CI

[folkertdev]

#1893 fixes the loongarch issue

[folkertdev]

In this context though, would it be appropriate?

Hmm, let's just leave it as Constraint for now then and I'll think about it some more.