Support for utils::globalVariables()

[MichalLauer]

If a function uses data masking, the function globals::globalsOf(...) throws an error, as expected.

example <- function() {
  mtcars |>
    dplyr::filter(cyl == 4)
}

globals::globalsOf(example, mustExist = T)
#> Error in globalsByName(names, envir = envir, mustExist = mustExist): Identified global objects via static code inspection (function (); {; dplyr::filter(mtcars, cyl == 4); }). Failed to locate global object in the relevant environments: 'cyl'

This can be fixed with two things: defining the variable as NULL

example <- function() {
  cyl <- NULL
  mtcars |>
    dplyr::filter(cyl == 4)
}

globals::globalsOf(example, mustExist = T)
# No error thrown...

or using pronouns.

example <- function() {
  mtcars |>
    dplyr::filter(rlang::.data$cyl == 4)
}

globals::globalsOf(example, mustExist = T)
# No error thrown...

However, when a global variable is defined using the utils::globalVariables function, the globals::globalsOf function throws an error.

example <- function() {
  utils::globalVariables(c("cyl"))

  mtcars |>
    dplyr::filter(cyl == 4)
}

globals::globalsOf(example, mustExist = T)
#> Error in globalsByName(names, envir = envir, mustExist = mustExist): Identified global objects via static code inspection (function (); {; utils::globalVariables(c("cyl")); dplyr::filter(mtcars, cyl == 4); }). Failed to locate global object in the relevant environments: 'cyl'

This is especially painful when developing R Shiny applications as a package.

I think adding this functionality would be a great addition, because is recommended in the book R Packages (2e) and in my humble opinion, it makes sense as the variable is defined globally.

What do you guys think? :)

[schloerke]

because is recommended in the book R Packages (2e)

From R Packages (2e)...

The “no visible binding” note is a peculiarity of using dplyr and unquoted variable names inside a package, where the use of bare variable names (english and temp) looks suspicious. You can add either of these lines to any file below R/ to eliminate this note

Using utils::globalVariables() is there to "eliminate this note". It does not actually set a global value that can be found.

---

Previous issue on finding globals given non-standard evaluation: #35

---

I don't believe utils::globalVariables() should be supported by {globals} even though the name implies "setting a globally available name".

[MichalLauer]

Hello @schloerke, understood. Thanks for the quick answer :)