Typed Column

[tyt2y3]

This is still WIP but I want everyone to see this earlier.

So basically after some research I figured we can have typed columns without the complexity of "one struct for every column", and this solution solves three things:

1. Make column methods type safe
2. No longer have to use CamelCase for column identifier
3. Make the shorthand for cake::Entity more useful

We would generate the following structs, columns share the same type so it's fine if Entity has 100 Columns:

pub struct TypedColumn {
    pub id: NumericColumn<Entity>,
    pub name: StringColumn<Entity>,
    pub price: NumericColumn<Entity>,
}

impl Entity {
    pub const COLUMN: TypedColumn = TypedColumn {
        id: NumericColumn(Column::Id),
        name: StringColumn(Column::Name),
        price: NumericColumn(Column::Price),
    };
}

// before
let cakes = cake::Entity::find()
    .filter(cake::Column::Name.like("Ch%"))
    .all(db)
    .await?;

// after
let cakes = Cake::find()
    .filter(Cake::COLUMN.name.like("Ch%"))
    .all(db)
    .await?;

The following would not compile:

Cake::find()
    .filter(Cake::COLUMN.id.like("Ch")) // method cannot be called on `NumericColumn<Entity>`
    .filter(Cake::COLUMN.name.like(2))  // the trait `From<{integer}>` is not implemented for `String`
    .order_by_desc(Cake::COLUMN.name)
    .one(db)
    .await?

This also means that contains for String would not collide with array_contains.

[tyt2y3]

Everything will pass after I bump sea-query

[tyt2y3]

we can finally have this

[tyt2y3]

release early, release often