RFC: Add a zero constant to quantities based on primitive types.

[adamreichold]

The public helper trait is a bit cumbersome but it appears to be the most straight-forward way to construct a zero value in const context.

An alternative that avoid the intermediate trait this would be to implement something like

impl<U> $quantity<U, $type>
where
    U: __system::Units<$type> + ?Sized,
{
    /// Constant representing the zero value.
    pub const ZERO: Self = Self {
        dimension: $crate::lib::marker::PhantomData,
        units: $crate::lib::marker::PhantomData,
        value: 0 as $type,
    };
}

for the same set of primitive types.

Fixes #26

[iliekturtles]

Thanks so much for the PR! I didn't intend to do a full review yesterday morning and completely forgot to thank you.

[iliekturtles]

Couple of minor things and I think we're good to merge.

Thoughts about the proposed new_in_base_units. Just push that into the const fn issue or create a new issue?

[adamreichold]

Thoughts about the proposed new_in_base_units. Just push that into the const fn issue or create a new issue?

I think a new issue is the better approach because our goal should probably be to have Quantity::new be a const fn even though this is currently still blocked on the trait bounds as well as const eval for floating point math. We could use a new issue to e.g. discuss adding nightly only features or macro-based workarounds.

[iliekturtles]

Thanks so much for this PR. I'll create a new issue to continue efforts for new / new_from_base_units