aboutsummaryrefslogtreecommitdiff
path: root/internal/config
diff options
context:
space:
mode:
authorLukasz Kasprzak <lukas@labunix.xyz>2026-08-18 16:02:17 +0200
committerLukasz Kasprzak <lukas@labunix.xyz>2026-08-18 16:02:17 +0200
commit27acee1930dc728b286c655cb66bd5e44804df6f (patch)
tree590651b88a173f6e35e2ebfa53d945a12b988425 /internal/config
parent967e911728eefb85dc53065c73c5c5d4cdce6a59 (diff)
downloadlectio-27acee1930dc728b286c655cb66bd5e44804df6f.tar.gz
lectio-27acee1930dc728b286c655cb66bd5e44804df6f.zip
fix(ef): Christmas Time runs to 13 January (RG 72)
RG 72: "Tempus natalicium decurrit a I Vesperis Nativitatis Domini usque ad diem 13 ianuarii INCLUSIVE." 6-13 January is the Epiphany section OF Christmas Time, not the start of Time after Epiphany. RG 119(a) backs the same boundary from the colour side, white "usque ad expletum tempus Epiphaniae". lectio ended Christmas Time on 5 January, so eight days a year came out green that should be white, and Epiphany itself reported the wrong season. NOT a one-line boundary change, which is why an earlier attempt was reverted rather than shipped. efSeason both named the slugs and reported the season, so moving the boundary renamed every slug in the window -- ef-time-after-epiphany-* became ef-christmas-* -- and orphaned the ef-time-after-epiphany-sunday-1 lectionary key. The two are now separate axes: efSlugSeason names slugs and numbers weeks; unchanged behaviour efSeason the liturgical season, which the colour follows; RG 72's The split is not a new idea in this file. The resumed-Sundays branch already does exactly this and says so -- "Season stays time-after-pentecost (calendrical); the Epiphany slug only routes the readings" -- it simply had no name, so the one place that needed it could not reuse it. Result, checked against colitur across all 365 days of 2026: SEASON now agrees on every single day, where it previously differed on eight. Colour differences drop to nine, every one a Saturday and every one the BVM Saturday Office (RG 78, colitur C17) which lectio does not build. Slugs and week numbers are byte-identical to before. The oracle's season axis now consults the allow-list and excludes excused days from its threshold. Neither was true before, because no cited divergence had ever BEEN a season difference -- the January window showed up as a colour one, since the wrong season produced the wrong colour too. With the season right, what is left is a season difference against missalemeum, and an axis whose threshold is a t.Fatalf would have failed regardless of any allow-list. TestTemporalEFChristmastideBoundary pins the boundary AND the split, mutation- tested both ways: reverting the boundary reddens it, and re-merging the two seasons so the slug is built from the liturgical one reddens it differently, naming the orphaned lectionary key.
Diffstat (limited to 'internal/config')
0 files changed, 0 insertions, 0 deletions