Fix single value detection in has_flipped_aes()

[yutannihilation]

These lines are to detect the case "the axes is a single value," but, in actual, they check whether all values are a fixed value1 or not.

ggplot2/R/utilities.r

Lines 594 to 600 in 5e6b1e5

	# Is one of the axes a single value
	if (all(x == 1)) {
	return(main_is_continuous)
	}
	if (all(y == 1)) {
	return(!main_is_continuous)
	}

This causes different behaviours between the case of 1 and that of another value. This PR fixes the check to zero_range(range()).

library(ggplot2)

df <- data.frame(value = 1:3, group = c("a", "a", "b"))
l <-  stat_summary(fun = "mean", geom = "bar", position = "dodge")
p1 <- ggplot(df, aes(value, 1, fill = group)) + l
p2 <- ggplot(df, aes(value, 0, fill = group)) + l

patchwork::wrap_plots(p1, p2, ncol = 1)

^{Created on 2020-01-29 by the reprex package (v0.3.0)}

[thomasp85]

The I tend of that check is really to catch when boxplots are created with no x-aesthetic at which point they’ll get a 1 automatically. I’ll have to think about the repercussions of this change a bit and whether it is the right way to go about it

[yutannihilation]

Got it, thanks for the context. Then, if this requires some time to consider, should I close this and resend against the master? This pull request is just that I was not immediately sure what the comment says.

[thomasp85]

Your reprex suggest that the check is too eager, not the other way around... I'll experiment a bit

[yutannihilation]

Your reprex suggest

To be clear, my reprex in the description shows the problematic behaviour of the current RC version. Sorry if it's confusing.

[thomasp85]

Closing in favour of #3800