Mastodawn

Just some guys

Rather happy with how well this worked

let fill_circ ~x ~y ~radius ?(r=0) ?(g=0) ?(b=0) screen =
    let t1 = ref (radius / 16) in
    let t2 = ref !t1 in
    let cx, cy = ref radius, ref 0 in
    while !cx >= !cy do
        for py = -(!cy) to !cy do
            set_pixel screen (!cx + x) (py + y) ~r ~g ~b ;
            set_pixel screen (-1 * !cx + x) (py + y) ~r ~g ~b ;
        done ;

        for py = -(!cx) to !cx do
            set_pixel screen (!cy + x) (py + y) ~r ~g ~b ;
            set_pixel screen (-1 * !cy + x) (py + y) ~r ~g ~b ;
        done ;

        cy := !cy + 1 ;
        t1 := !t1 + !cy ;
        t2 := !t1 - !cx ;
        if !t2 >= 0 then begin
            t1 := !t2 ;
            cx := !cx - 1 ;
        end
    done

Show thread

or, what?2d ago

It's nice when your observations were in fact correct and also easy to implement.

Show thread

mx alex tax1a - 2020 (6)2d ago

@andnull tbh im an advocate of tasteful application of mutable state

Show thread

or, what?2d ago

@atax1a I am so thankful OCaml respects the need for mutable state. Having to contort everything into recursive iteration can be such a pain in the fuckin ass. Especially, when you are standing directly on the FFI boundary.

This whole little project has been "purely procedural" OCaml and it's extremely nice to use. There is very much a good reason to use OCaml even if you aren't leveraging it's functional origins.

Show thread

or, what?

@atax1a What really tells me OCaml gets it is the incluse of downto instead of some clumsy "iterate with a negative step" bullshit that like... literally every procedural language asks you to do...

Show thread

mx alex tax1a - 2020 (6)2d ago

@andnull downto is the, like, one Pascal feature we don't find to be arrant european puffery, so, yeah