#2611 Fix week() calculation to start week 1 at days 1–7

[venom1204]

Closes #2611

Hi @jangorecki, @tdhock, and @ben-schwen,

Due to some issues, I had to delete the previous branch and continue the work in this new one.

Thank you for the earlier feedback! I’ve updated the PR to address the concerns regarding the breaking change in week().

The updated fix now includes the warning system as discussed:

• The warning is shown only when the result differs from the old (buggy) behavior.
• It appears only once per session.
• It can be suppressed entirely using options(datatable.warn.week.change = FALSE).

I’ve also updated NEWS.md and the test suite to reflect these changes.
This should now cover all the requested adjustments — does this modification align with what you had in mind?

If any further tweaks or changes are needed, please let me know.

Thanks again for your time and guidance!

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.13%. Comparing base (b3acf2a) to head (eff4fad).
⚠️ Report is 2 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7380   +/-   ##
=======================================
  Coverage   99.13%   99.13%           
=======================================
  Files          85       85           
  Lines       16556    16577   +21     
=======================================
+ Hits        16412    16433   +21     
  Misses        144      144           

☔ 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.

[github-actions]

• HEAD=issue_2611 slower P<0.001 for memrecycle regression fixed in #5463
  

Generated via commit eff4fad

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 36 seconds

[jangorecki]

How about warning about coming fix in next release now, and switching the fix on in next dev cycle? And possibly option to force new behavior already now, so revdep code can be already adapted.
Change like this can literally cost money for automates decision systems.

[jangorecki]

Sorry for closing, it must have been an accident.

[ben-schwen]

last two style NITs and then we are good go to go from my side!

NB: when we do suggestions for a single test like the options style we suspect all tests to adapt to this style (otherwise we would introduce inconsistencies which reads even worse than different styles)

[ben-schwen]

TY!