[dajiaohuang]

Summary of changes

{Wmon}, {Wsun}, %W and %U return values that are off by one week in years beginning on a Sunday or Monday, and treat the days before the first weekstart day of the year as week 1 instead of week 0. This is the regression reported in #742; it was introduced when the week number calculation was reworked.

The previous implementation anchored the week boundary at the start of the year and then divided the shifted day number by 7, which only happens to be correct when the year starts on the weekstart day. The new implementation derives the week number from the day of the year and the weekday offset directly, so that days before the first weekstart day belong to week 0 and the following weeks are numbered from there, as strftime defines %W and %U.

Also corrects the documented range of these directives from 01..53/00..52 to 00..53.

Fixes #742

Checklist

• New functions have typespecs, changed functions were updated
• Same for documentation, including moduledocs
• Tests were added or updated to cover changes
• Commits were squashed into a single coherent commit
• Notes added to CHANGELOG file which describe changes at a high-level

[bitwalker]

Thanks for the contribution! The week-numbering fix is now on main in be1641bf as part of the Timex 4.0.0 changes.

{Wmon}, {Wsun}, %W and %U now start week one at the first Monday or Sunday, with week zero for preceding days. Regression coverage includes the affected formats and year boundaries.

Closing as superseded by that implementation.