guides() breaks old guides if the first argument is an old guide

[eliocamp]

The old guide system used lists. This trips up a check in guides(), which throws this spurious warning:

library(ggplot2)

old_guide <- metR:::guide_colourstrip()
is.list(old_guide)
#> [1] TRUE

new_guide <- guide_colorbar()
is.list(new_guide)
#> [1] FALSE

mtcars |> 
    ggplot(aes(gear, carb)) +
    geom_point() +
    guides(fill = old_guide)
#> Warning: The `<scale>` argument of `guides()` cannot be `FALSE`. Use "none" instead as
#> of ggplot2 3.3.4.
#> This warning is displayed once every 8 hours.
#> Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
#> generated.

The problem is here:

ggplot2/R/guides-.R

Lines 76 to 83 in 97edd62

	if (is.list(args[[1]]) && !inherits(args[[1]], "guide")) args <- args[[1]]
	args <- rename_aes(args)
	idx_false <- vapply(args, isFALSE, FUN.VALUE = logical(1L))
	if (isTRUE(any(idx_false))) {
	deprecate_warn0("3.3.4", "guides(`<scale>` = 'cannot be `FALSE`. Use \"none\" instead')")
	args[idx_false] <- "none"
	}

Line 76 seems to be there to support a guides(list(fill = ..., colour = ...)) syntax. But when using the old guide system, the logic makes args point to the first guide. Then line 79 iterates over the elements of the guide and if one of them is FALSE then it converts it to "none" and throws the warning.

[teunbrand]

Thanks for the report Elio! I think this might be a duplicate of #6165, for which we merged a fix 3 days ago. Does updating the development version of ggplot2 help at all?