Fixed inconsistencies in .ci/atime/tests.R

[Anirban166]

I noticed some minor inconsistencies in the file, for e.g. double expr from #6288: (also, using curly braces for only multiple lines might be good for consistency)

data.table/.ci/atime/tests.R

Lines 137 to 139 in 687de9d

	expr = {
	expr=data.table:::`[.data.table`(d, , max(v1) - min(v2), by = id3)
	},

Originally posted by @Anirban166 in #6464 (comment)

Also, some comments needed corrections, such as:

data.table/.ci/atime/tests.R

Line 124 in 365d2a3

Fast = "1872f473b20fdcddc5c1b35d79fe9229cd9a1d15"), # Last commit in the PR that fixes the issue (https://github.com/Rdatatable/data.table/pull/5427/commits)

(third-last commit, not last)

[github-actions]

Generated via commit ef88d6f

Download link for the artifact containing the test results: ↓ atime-results.zip

Time taken to finish the standard R installation steps: 3 minutes and 28 seconds

Time taken to run atime::atime_pkg on the tests: 6 minutes and 35 seconds

[Anirban166]

@MichaelChirico @tdhock Seems like the setup time decrease is worthwhile/significant! :D
(This is showing results that are consistent with what I'm getting on my fork now, for e.g. the first run there was v1.2.0 taking ~ 12 minutes for setup vs v1.2.1 now taking just 3 and a half minutes)

[tdhock]

I don't understand Antepenultimate, why did you change from the simpler "last" ?

[Anirban166]

The SHA is not associated with the last commit (you can check the link and verify), but I agree that it's better to avoid such a jargon - will 'third-last' or 'close-to-last' work?

[MichaelChirico]

Is there something wrong with the last SHA in the PR, then just "last commit" will be correct? otherwise, third-to-last seems best to me (as much as I hate wasting such a perfectly valid word!)

[Anirban166]

Great point! No issues with the SHA, just didn't get updated at the end of #6094 unfortunately (I should've tried it with the last one before changing the comment)

Thank you both!