fix(ef): weeks after Epiphany count from the first Sunday, not from Epiphany

efWeek anchored the Time-after-Epiphany week number at 6 January, which put
every week of the season one too high -- 14 January 2026 came out as week 2
when the first Sunday after Epiphany was 11 January and the week running from
it is the first. The error persisted from Epiphany until Septuagesima cut the
season short, so it was wrong for most of January and February every year.

The Sunday is the anchor because it carries the Mass the week is named for,
"Dominica I post Epiphaniam", and its ferias follow it. Epiphany itself is a
feast inside Christmas Time (RG 72), not the head of a numbered week -- which
is the same distinction the Christmas Time boundary fix just made, showing up
again one function along.

Week numbers now match colitur day for day through the season.

The test pins the ordinary years AND the edge that a naive fix gets wrong:
when Epiphany itself falls on a Sunday, "the first Sunday AFTER Epiphany" is
the following week, so the helper starts its search on 7 January rather than
the 6th. It finds such a year in 2005..2050 rather than hardcoding one.
Mutation-tested: restoring the 6 January anchor reddens it.

0 files changed, 0 insertions, 0 deletions