Should env_ functions be internally generic

[hadley]

Not the parent of that environment.

library(rlang)

f <- function() {}
env_parent(f)
#> <environment: package:rlang>
#> attr(,"name")
#> [1] "package:rlang"
#> attr(,"path")
#> [1] "/Users/hadley/R/rlang"

Created on 2018-02-28 by the reprex package
(0.1.1.9000).

[lionel-]

Should we deprecate the internal genericity of env_ functions?

[hadley]

I renamed the title of this issue. Another typo that this behaviour makes confusing:

library(rlang)
env_bind(env, x = 1)
#> Error in base::assign(nm, data[[nm]], envir = env_): cannot add bindings to a locked environment

Created on 2018-03-08 by the reprex package (v0.2.0).