Feature request: label_log() but for negative numbers

[teunbrand]

Some distributions are logarithm-like, like inverse hyperbolic sine (asinh) and pseudo-logarithmic transformations, but they can handle negative numbers.
However, label_log() does not handle the labelling very gracefully for these transformations.

library(scales)

demo_continuous(
  c(-100, 100),
  transform = "asinh",
  breaks = c(-100, -10, -1, 0, 1, 10, 100),
  labels = label_log()
)
#> scale_x_continuous(transform = "asinh", breaks = c(-100, -10, 
#>     -1, 0, 1, 10, 100), labels = label_log())
#> Warning in format(log(x, base = base), digits = digits): NaNs produced

^{Created on 2024-04-06 with reprex v2.1.0}

The ask here is to either include an option in label_log() to handle signed numbers or have label_log() handle signed numbers by default. I'm proposing the plot above shows something like the plot below:

library(scales)

demo_continuous(
  c(-100, 100),
  transform = "asinh",
  breaks = c(-100, -10, -1, 0, 1, 10, 100),
  labels = expression(-10^2, -10^1, -10^0, 0, +10^0, +10^1, +10^2)
)
#> scale_x_continuous(transform = "asinh", breaks = c(-100, -10, 
#>     -1, 0, 1, 10, 100), labels = expression(-10^2, -10^1, -10^0, 
#>     0, +10^0, +10^1, +10^2))

^{Created on 2024-04-06 with reprex v2.1.0}