Normally I would think curved in plan. OTOH, if a greater part of the spoiler is more oblique than a given curved plan, as in this case were the forward proximity isnt far from the most rearward, then a V may be better, and will avoid a high pressure bubble which could tend to lift the front of the car, something that true air dams do not endure (being at the extreme front).
You can resolve this by testing two plan examples against each other, and see what proportion of the parabolic curve is steeper than the diagonal by taking test samples of departure angle at even stations across the width.
Using straight geometry you will quickly see how the proximity of your most forward point plays this out, hard to tell in one dimension
I think it needs a lip as well