Bare type predicates: explicitly call out matrices&arrays in docs (as well as names & other attributes?)

[brookslogan]

The current documentation for the bare type predicates includes:

#' These predicates check for a given type but only return `TRUE` for
#' bare R objects. Bare objects have no class attributes. For example,
#' a data frame is a list, but not a bare list.

Aside from what appears to be a minor typo ("attributes"), this appears to be technically accurate, but likely surprising to many in the case of matrices and arrays:

mat <- matrix(1:6, 2, 3)
ray <- array(1:12, c(2, 2, 3))
rlang::is_bare_numeric(mat)
#> [1] TRUE
rlang::is_bare_numeric(ray)
#> [1] TRUE
names(attributes(mat))
#> [1] "dim"
names(attributes(ray))
#> [1] "dim"
class(mat)
#> [1] "matrix" "array"
class(ray)
#> [1] "array"

^{Created on 2025-02-04 with reprex v2.1.1}

It would be helpful to mention this case in the documentation. It may also help to explicitly mention that other attributes may still be present. E.g., maybe something like:

#' These predicates check for a given type but only return `TRUE` for
#' bare R objects. Bare objects have no class attribute. For example,
#' a data frame is a list, but not a bare list. Bare objects include matrices
#' and arrays unless they have been assigned a non-`"matrix"`, non-`"array"`
#' class. They can also have non-class attributes such as names.

though it's a bit wordy. Or maybe something like

#' Bare objects include many matrices and arrays; see [`class`]."