Improve translation-readiness of string with subclause

[MichaelChirico]

Closes #7475

@rffontenelle maybe this could be %s resolving to in serial or in parallel? I am unclear if it's as easy as swapping the word out in other languages (probably @aitap has the best guess 😄 )

[github-actions]

• HEAD=i18n-msg-split slower P<0.001 for DT[by,verbose=TRUE] improved in #6296
  

Generated via commit ba4e812

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

Task	Duration
R setup and installing dependencies	2 minutes and 51 seconds
Installing different package versions	43 seconds
Running and plotting the test cases	2 minutes and 55 seconds

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.04%. Comparing base (6329d94) to head (ba4e812).
⚠️ Report is 16 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #7476      +/-   ##
==========================================
- Coverage   99.06%   99.04%   -0.03%     
==========================================
  Files          86       87       +1     
  Lines       16616    16678      +62     
==========================================
+ Hits        16461    16518      +57     
- Misses        155      160       +5     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[aitap]

This particular sentence might actually work even if translated from small parts. Since the %s part is in parentheses, we can pretend that it's a small sentence. If we really wanted to build sentences from small parts, we could try contexts so that two different occurrences of in parallel could be translated differently if needed, but duplication as currently in this PR is more in line with how gettext is intended to be used.

[jangorecki]

I would use "sequentially" rather than "in serial", but that's a very minor thing so LGTM

[MichaelChirico]

Ah, ideally this would have waiting until 1.18.0 was fully on CRAN 😬

Now we have this string in .pot that's no longer in the package:

data.table/po/data.table.pot

Lines 2570 to 2573 in e455714

	#: froll.c:1628
	#, c-format
	msgid "%s: finding order and initializing links for %d blocks %stook %.3fs\n"
	msgstr ""