`coord_cartesian()` should probably throw an error is `xlim` or `ylim` is length 1

[DavisVaughan]

I accidentally ran this, and was rather confused for awhile until I realized I hadn't supplied an upper limit in xlim:

library(ggplot2)

df <- data.frame(
  x = 1:5,
  y = 2:6
)

ggplot(df, aes(x, y)) +
  geom_line() +
  coord_cartesian(xlim = 2)
#> Error in if (all(is.finite(continuous_range_coord)) && diff(continuous_range_coord) < : missing value where TRUE/FALSE needed

Could coord_cartesian(), or possibly something later on, throw an error if it doesn't get a length 2 xlim / ylim?

[teunbrand]

Thanks Davis! A similar suggestion was made here: #4894 (comment), so hopefully this PR will fix this.

[DavisVaughan]

I don't think is.vector() is a particularly good helper to check for this

is.vector(x) returns TRUE if x is a vector of the specified mode having no attributes other than names.

Meaning that the limits can't be any custom data types (since if a class attribute exists it will return false)

[teunbrand]

You're right, might be simplified to vctrs::vec_check_size() perhaps.

[teunbrand]

I'm gonna mark this as completed in #4894.