[R] Two tests are failing with to-be-released dplyr dependency

[paleolimbot]

Describe the bug, including details regarding any error messages, version, and platform.

Error (test-dplyr-join.R:241): semi_join
<rlib_error_dots_nonempty/rlib_error_dots/rlang_error/error/condition>
Error in `semi_join(., to_join, by = "some_grouping", keep = TRUE)`: `...` must be empty.
x Problematic argument:
* keep = TRUE
Backtrace:
 1. arrow:::compare_dplyr_binding(...)
      at test-dplyr-join.R:241:2
 6. dplyr:::semi_join.data.frame(., to_join, by = "some_grouping", keep = TRUE)

(Looks like we just need to remove keep = TRUE, which was never an argument to semi_join() anyway?)

Error (test-dplyr-funcs-datetime.R:3641): with_tz() and force_tz() works
<dplyr:::mutate_error/rlang_error/error/condition>
Error in `mutate(., timestamps_with_tz_1 = force_tz(timestamps, "Europe/Brussels", 
    roll_dst = c("boundary", "post")))`: i In argument: `timestamps_with_tz_1 = force_tz(...)`.
Caused by error in `force_tz()`:
! unused argument (roll_dst = c("boundary", "post"))
Backtrace:
  1. arrow:::compare_dplyr_binding(...)
       at test-dplyr-funcs-datetime.R:3641:2
  6. dplyr:::mutate.data.frame(...)
  7. dplyr:::mutate_cols(.data, dplyr_quosures(...), by)
  9. dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
 10. mask$eval_all_mutate(quo)
 11. dplyr (local) eval()

Component(s)

R

[assignUser]

Should this be a blocker? We could cherry pick it into the cran release but eh

[paleolimbot]

It would be nice to not have to cherry pick! I think this one is a quick.

There's also an LTO nightly failure that's probably an arrow C++ problem (also probably a quick fix).

[assignUser]

Marked as blocker as worst case it will necessitate another CRAN release

[thisisnic]

@paleolimbot Where did you spot the second test failure? I couldn't reproduce it locally after installing dev dplyr.

[paleolimbot]

I believe that's because I hadn't installed the latest lubridate on my laptop. We don't have to have our tests pass for outdated lubridate so feel free to ignore it!