MINOR: [R] Remove outdated TODO for integer division by zero tests

[HyukjinKwon]

Rationale for this change

63b8b7e: ARROW-13800 Added the TODOs
2468992: ARROW-14297 Added the tests but forgot to remove TODOs.

arrow/r/tests/testthat/test-compute-arith.R

Line 67 in 2468992

expect_equal(a %/% 0, Array$create(c(Inf, Inf, Inf, Inf, NA_real_)))

arrow/r/tests/testthat/test-compute-arith.R

Line 70 in 2468992

expect_equal(a %/% 0L, Array$create(rep(NA_integer_, 5)))

arrow/r/tests/testthat/test-compute-arith.R

Lines 79 to 80 in 2468992

	expect_equal(b %/% 0, Array$create(c(Inf, Inf, Inf, Inf, NA_real_)))
	expect_equal(b %/% 0L, Array$create(c(Inf, Inf, Inf, Inf, NA_real_)))

What changes are included in this PR?

This PR proposes to remove the obsolete todo about integer division by zero tests.

Are these changes tested?

No, I did not test this.

Are there any user-facing changes?

No, dev-only.

[jonkeane]

Thanks for this! One small comment, I don't mind duplicating the test behavior in the place where the todo was so that we have all the arithmetic tests in one place (even though they are also tested in the dplyr tests too, we already have a bunch of overlap there too)

[jonkeane]

I wouldn't mind also having a test here as well as the dplyr tests. Or alternatively: maybe we mention that the %/% tests are in the dplyr math file?

[HyukjinKwon]

Oh actually, I believe they are already being tested above:

...
  expect_equal(a %/% 0, Array$create(c(Inf, Inf, Inf, Inf, NA_real_)))
...
  expect_equal(a %/% 0L, Array$create(rep(NA_integer_, 5)))
...
  expect_equal(b %/% 0, Array$create(c(Inf, Inf, Inf, Inf, NA_real_)))
  expect_equal(b %/% 0L, Array$create(c(Inf, Inf, Inf, Inf, NA_real_)))
...

2468992 also added the zero tests into here. Let me fix the PR description!

[jonkeane]

Aaah good catch! Thanks for the cleanup

[HyukjinKwon]

gentle ping on this @jonkeane 🫡

[jonkeane]

Sorry for the delay!

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit a0f3d7b.

There weren't enough matching historic benchmark results to make a call on whether there were regressions.

The full Conbench report has more details.