[FR] Allow showing expression instead of values in arg_match() error message

[mikmart]

Sometimes I need to match an argument against a long (100+) list of valid values. Currently the arg_match() error message will be overwhelmed by listing out all possible valid values. In such cases, it would be useful to have a way to include the expression from the formal argument in the message instead.

For example with:

foo <- function(dataset = ls("package:datasets")) {
  rlang::arg_match(dataset)
}

foo("Titanci")
#> Error in `foo()`:
#> ! `dataset` must be one of "ability.cov", "airmiles", "AirPassengers",
#>   "airquality", "anscombe", "attenu", "attitude", "austres", "beaver1",
#>   "beaver2", "BJsales", "BJsales.lead", "BOD", "cars", "ChickWeight",
#>   "chickwts", "co2", "CO2", "crimtab", "discoveries", "DNase", "esoph", "euro",
#>   "euro.cross", "eurodist", "EuStockMarkets", "faithful", "fdeaths",
#>   "Formaldehyde", "freeny", "freeny.x", "freeny.y", "HairEyeColor",
#>   "Harman23.cor", "Harman74.cor", "Indometh", "infert", "InsectSprays", "iris",
#>   "iris3", "islands", "JohnsonJohnson", "LakeHuron", "ldeaths", "lh",
#>   "LifeCycleSavings", "Loblolly", "longley", "lynx", "mdeaths", "morley",
#>   "mtcars", "nhtemp", "Nile", "nottem", "npk", "occupationalStatus", "Orange",
#>   "OrchardSprays", "PlantGrowth", "precip", "presidents", "pressure",
#>   "Puromycin", "quakes", "randu", "rivers", "rock", "Seatbelts", "sleep",
#>   "stack.loss", "stack.x", "stackloss", "state.abb", "state.area",
#>   "state.center", "state.division", "state.name", "state.region", "state.x77",
#>   "sunspot.month", "sunspot.year", "sunspots", "swiss", "Theoph", "Titanic",
#>   "ToothGrowth", "treering", "trees", "UCBAdmissions", "UKDriverDeaths",
#>   "UKgas", "USAccDeaths", "USArrests", "UScitiesD", "USJudgeRatings",
#>   "USPersonalExpenditure", "uspop", "VADeaths", "volcano", "warpbreaks",
#>   "women", "WorldPhones", or "WWWusage", not "Titanci".
#> ℹ Did you mean "Titanic"?

It would be nice to be able to instead get:

foo("Titanci")
#> Error in `foo()`:
#> ! `dataset` must be one of `ls("package:datasets")`, not "Titanci".
#> ℹ Did you mean "Titanic"?