expect_lint() error for missing {testthat} should reference which function called it

[MichaelChirico]

          ideally this will just reference the actual call for clarity, but we can save that for later.

Originally posted by @MichaelChirico in #2582 (comment)

[F-Noelle]

I am thinking something along the line of:

who_called_me <- as.character(sys.call(sys.parent())) |>
  stringr::str_split(stringr::fixed("("))
who_called_me <- who_called_me[[1]]

I am not sure if there would be a more elegant way to get this info which function did the original call.

Or you create a third unexported function called "expect_lint_logic" who does all the heavy lifting and has an addition parameter which is simply the name of the calling function.

Any better ideas?

I wanted to write the expect_no_lint_linter as well, given that this should be a fairly easy linter to write as it just needs to check for NULL on the second positional parameter of the function, as far as I understood the xpath logic

[MichaelChirico]

I wanted to write the expect_no_lint_linter as well, given that this should be a fairly easy linter to write as it just needs to check for NULL on the second positional parameter of the function, as far as I understood the xpath logic

Follow #1874 for some discussion about how exactly to go about that. We might want something more generalizable than just expect_no_lint_linter().

[AshesITR]

Regarding the caller logic, you can look at the code of Linter() which auto-names using a similar approach.

[MichaelChirico]

I would use:

tail(as.character(sys.call(sys.parent())[[1]]), 1L)

That covers lintr::expect_no_lint() and plain expect_no_lint().

The lintr:::linter_auto_name() logic might be overkill in this case.

Alternatively, we can just copy-paste the requireNamespace() call into the body of expect_no_lint().

[AshesITR]

Alternatively, we can just copy-paste the requireNamespace() call into the body of expect_no_lint().

This is the way to go I think.

[F-Noelle]

Probably the most simple solution indeed.