feature request: geom_arc_bar with categorical axis

[romainFr]

For now geom_arc_bar only understands x0 and y0 as numeric (on a continuous axis). Would it be too strange to have it working with categorical axis?
Minimum example:

catpie <- data.frame(x=c("A","B","A","C"),y=c("1","2","2","1"),v1=c(0.5,0.8,0.6,1))
ggplot(catpie) + geom_arc_bar(aes(x0=x,y0=y,r0=0,r=0.5,amount=v1),stat="pie") + scale_x_discrete()

Error: Discrete value supplied to continuous scale
In addition: Warning messages:
1: In Ops.factor(paths$x, paths$x0) : ‘+’ not meaningful for factors
2: In Ops.factor(paths$y, paths$y0) : ‘+’ not meaningful for factors

And thanks for all the great work !

[thomasp85]

I wouldn't really know how to interpret a discrete scale in the context of an arc... arcs exist in 2D Euclidean space and thus requires two numerical dimensions to make sense

[funnell]

@thomasp85 wouldn't the ordered discrete axis elements just be mapped to a numerical scale, practically?
For what it's worth, I'd also love to be able to provide categorical variables to set x and y positions of arcs.