fix(lectionary): two Judith citations wrong in both engines

A live-missalemeum sweep over a full liturgical year (2028, 364 days)
returned 333 matching and 31 differing days: 17 already-known or pure
notation, 3 new. All three now adjudicated against the photographic
scans of the 1962 Missal.

The Assumption's Epistle reads "Iudith 13, 22-25; 15, 10" (scan1:35357).
Both engines carried "Judith 13:22-25; 13:15; 13:10" -- a chapter
carry-over defect that re-applied chapter 13 to the second citation and
so invented two verses the Missal never cites. The Seven Sorrows reads
"Iudith 13, 22 et 23-25" (scan1:36938, corroborated at 28468); both
carried "Judith 13:22; 13:25", dropping the range. The third, St Monica,
resolves the other way: the Missal gives 1 Timothy (scan2:32679, heading
and incipit both legible; scan1:29524 agrees) where missalemeum serves
the Common of Holy Women's Proverbs -- colitur is right, no change.

Layer 3 could not have found either citation defect. colitur's data is
bootstrapped from lectio, and lectio is the other side of that
differential, so while both engines produced the identical wrong answer
the comparison held it against itself and reported agreement -- the
shared-lineage blindness CLAUDE.md names, now with a second live witness
beside Holy Thursday's colour. Only an external oracle could see it, and
only outside the committed fixture's own window.

Fixed upstream in lectio, re-bootstrapped here (provenance SHA-256 moved
to 54f6bfc5, verified against the source), fixture regenerated.

The upstream fix inverted layer 3 from blind to load-bearing here:
reverting the Assumption citation now reddens both layers, where before
it would have reddened neither. Verified by mutation, both directions.

Allow-lists:
- C22 closed and removed (128 rows -> 0), not re-adjudicated -- the other
  engine was fixed to agree, the same shape as the C2-C5/C7/C9-C13/C21
  closures. Its own note had already flagged the Lenten Ember dead-key
  bug for upstream; lectio now also fails its build on any unreachable
  lectionary key rather than serving a plausible wrong Mass.
- M28 opened (3 rows, verdict colitur), pinning both sides' citation
  text rather than the date and diff shape alone, so a different
  divergence landing on the same date cannot be swallowed.
- M8's gate widened, expected_rows unchanged at 2: 2027-08-15 carries
  both its cause and M28's, and one entry is admitted per day. It stays
  under M8, which now requires its own commemoration difference to be
  present and admits the citation difference only when it satisfies
  M28's two-sided predicate. Neither cause absorbs the other.

layer_m_reason's oracle-row parameter, kept named since M2 closed for
exactly such a future reader, is live again.

Register: docs/research/rules-register.md section 6.6.

1 files changed, 7 insertions, 1 deletions

diff --git a/data/ef/expected-divergences-missalemeum.sexp b/data/ef/expected-divergences-missalemeum.sexp
index b5374ab..90ffceb 100644
--- a/data/ef/expected-divergences-missalemeum.sexp
+++ b/data/ef/expected-divergences-missalemeum.sexp
@@ -196,8 +196,14 @@ WIDENED, Task 9 (branch ef-lectionary, layer-4 oracle): [First_mismatch]/[Gospel
  ((id M8)
  (citation "RG 109(a) (\"of a Sunday\" is always privileged) + RG 111(a) (\"I class: none save one privileged\")")
  (verdict colitur)
- (note "When a fixed I-class feast (All Saints 1 Nov; the Assumption 15 Aug) lands on an ordinary Sunday, the impeded Sunday IS the one privileged commemoration RG 111(a) admits -- already an established, independently-tested rule (test_precedence_ef.ml's RG95/RG109(a) row). missalemeum shows no commemoration at all on either occurrence in this window.")
+ (note "When a fixed I-class feast (All Saints 1 Nov; the Assumption 15 Aug) lands on an ordinary Sunday, the impeded Sunday IS the one privileged commemoration RG 111(a) admits -- already an established, independently-tested rule (test_precedence_ef.ml's RG95/RG109(a) row). missalemeum shows no commemoration at all on either occurrence in this window. WIDENED, Task 9 (branch ef-lectionary): one of this entry's two dates, 2027-08-15, now ALSO carries M28's Judith-citation divergence -- the Assumption falls on a Sunday that year, so BOTH causes land on the identical day, and the classifier admits one entry per day. That day stays here rather than moving to M28, because THIS entry's cause is the one that decides the day's own commemoration; the gate requires [Comm_presence] to be actually present and admits [First_mismatch] only when it is exactly M28's shape with both sides pinned (test_oracle.ml's own [judith_divergence]), so neither cause is silently absorbed into the other and a citation divergence of any other shape on either M8 date still escapes to unexplained. expected_rows is unchanged at 2 for that reason; M28 declares the remaining 3 dates.")
  (expected_rows 2))
+
+ ((id M28)
+ (citation "MR1962 propers, primary-source-verified against the photographic scans (branch ef-lectionary, 2026-08-17): the Assumption at scan1.txt:35357, under \"IN ASSUMPTIONE BEATAE MARIAE VIRGINIS\" -- \"Iudith 13, 22-25; 15, 10\"; the Seven Sorrows at scan1.txt:36938, corroborated independently at scan1.txt:28468 -- \"Iudith 13, 22 et 23-25\".")
+ (verdict colitur)
+ (note "Two Judith citations both engines used to get wrong the same way, found by a live-missalemeum full-year sweep (2028 liturgical year) run this task, then adjudicated against the scans. lectio's generator emitted \"Judith 13:22-25; 13:15; 13:10\" for the Assumption -- a chapter carry-over defect, re-applying chapter 13 to the second citation's own \"15, 10\" and thereby INVENTING two verses -- and \"Judith 13:22; 13:25\" for the Seven Sorrows, dropping the \"et 23-\" range. colitur inherited both through its own bootstrap. This is the \"Holy Thursday was violet in both\" shared-lineage shape this project's CLAUDE.md names, and precisely why layer 3 could never have caught it: lectio IS the other side of that comparison, so the differential was structurally blind for as long as both engines agreed on the wrong answer. Fixed upstream in lectio and re-bootstrapped here; missalemeum still serves the defective form, so the divergence surfaces on layer 4 and is adjudicated colitur against the primary source. Gated on the literal date list AND on BOTH sides' citation text (test_oracle.ml's own [judith_divergence]) -- not merely the diff shape, which is the C6/C14 failure mode this file's header documents. 3 rows, not 4: the fourth affected date, 2027-08-15, is accounted under M8, which see.")
+ (expected_rows 3))
  ((id M10)
  (citation "RG 109(e), unqualified text: \"de feriis Adventus, Quadragesimae et Passionis\" -- of the FERIAS of Advent (not \"the Advent ferias 17-23 December only\", which is RG 91 entry 18's separate OCCURRENCE-table dignity, not a RG 109 privilege restriction)")
  (verdict colitur)