Using 'label_number' at package top-level generates R CMD check NOTE about '...'

[MichaelChirico]

foo <- scales::label_number(
  scale_cut = scales::cut_short_scale(),
  accuracy = 0.1
)

codetools::checkUsage(foo)
# <anonymous>: ... may be used in an incorrect context: 
#      number(x, accuracy = accuracy, scale = scale, prefix = prefix, 
#          suffix = suffix, big.mark = big.mark, decimal.mark = decimal.mark, 
#          style_positive = style_positive, style_negative = style_negative, 
#          scale_cut = scale_cut, trim = trim, ...)

Is that working as intended? Or is there a known workaround?

[teunbrand]

What does WAI mean?

[MichaelChirico]

Sorry for the jargon. Edited original -- it's 'working as intended'. I'll hide this soon to reduce noise.

[teunbrand]

The function does work as intended: ... arguments in the function factory are forwarded to base::format() when used to label numbers.

foo <- scales::label_number(
  scale_cut = scales::cut_short_scale(),
  accuracy = 0.1, 
  width = 7 # to base::format
)
foo(c(1, 10, 100)) |> 
  nchar()
#> [1] 7 7 7

^{Created on 2025-10-13 with reprex v2.1.1}

I'd agree that the codetools::checkUsage() outcome is unfortunate and it may be worth rephrasing the function to prevent the note.

[MichaelChirico]

I don't understand well why {codetools} is complaining, maybe it's a false positive?

[teunbrand]

Yes that seems right to my mind. The 'may' part of the complaint is the operative word, but I can understand that from a static analysis of foo's function body/formals the use of ... may be flagged.

[MichaelChirico]

I do think this is a {codetools} false positive:

https://gitlab.com/luke-tierney/codetools/-/issues/20