They are incompatible, and it makes it hard to combine grids of different layouts (see #217).
I'm trying to make them compatible. I have two hypotheses:
If LayoutPick and PickFn are on the same level of abstraction, then why they are not compatible? I don't see an easy way to convert between a function and a structure.
If they are entities of different "kinds", then LayoutPick may be wrapped in PickFn with const . Just, but I don't think this is a proper way, because this leads to PickFn (PickFn (LayoutPick _)) in rendering functions.
When I use functions like this:
renderableToFile def "/tmp/a.png"
:: Renderable a -> IO (PickFn a)
renderableToFile def "/tmp/a.png" $ gridToRenderable $ layoutToGrid def
:: (PlotValue x, PlotValue y) =>
IO (PickFn (LayoutPick x y y))
it looks like they are on different levels, and Renderable expects the argument to be LayoutPick to wrap it then in PickFn.
But when I use
pieChartToRenderable :: forall a. PieChart -> Renderable (PickFn a)
it happens that Renderable expects its argument to be PickFn! To wrap it into what?
They are incompatible, and it makes it hard to combine grids of different layouts (see #217).
I'm trying to make them compatible. I have two hypotheses:
If LayoutPick and PickFn are on the same level of abstraction, then why they are not compatible? I don't see an easy way to convert between a function and a structure.
If they are entities of different "kinds", then LayoutPick may be wrapped in PickFn with
const . Just, but I don't think this is a proper way, because this leads toPickFn (PickFn (LayoutPick _))in rendering functions.When I use functions like this:
it looks like they are on different levels, and Renderable expects the argument to be LayoutPick to wrap it then in PickFn.
But when I use
it happens that Renderable expects its argument to be PickFn! To wrap it into what?
Double PickFn?