feat(ef): the Gloria in excelsis, RG 431-432, deferring to Breviary 237-238

Phase 2 of celebrant-rubrics-phase1: colitur rubrics gains a fifth
column, whether the Gloria is said. Follows the Creed's own seam
exactly -- Rite.t.gloria, Liturgical_day.t.gloria, wired through
calendar.ml the same way.

RG 431(a)/432(a) defer the Gloria to the Breviary's own Te Deum rule
(nn. 237-238), so te_deum is implemented as its own named predicate,
cited clause by clause, not collapsed into a colour heuristic. 431(c)
(Holy Thursday, the Easter Vigil Mass) and 432(b)/(d) (violet; a
Requiem) are independent overrides checked ahead of the Te Deum-derived
answer. Every clause this engine has no dimension to model
(votive Mass classes, the wider n.302 "Missa festiva" categories) is
stated as N/A with its own reasoning, not silently dropped.

Validated against the FIUV universal Ordo (Gloria and Te Deum) and all
three LMS editions (Gloria). A first pass over-trusted a clean-looking
15-for-15 FIUV contradiction of 237(b)'s own Septuagesima exception and
replaced it with a blanket "every Sunday" rule; the evidence was itself
corrupted -- the FIUV extractor recognised only one of the source's two
Te Deum negations ("non dicitur", not "sine"), so every "sine Te Deum"
Sunday read wrongly true. Fixed in tools/extract_fiuv_ordo.ml, fixture
re-extracted, and the literal 237(b) reading restored once the
corrected data confirmed it. A second bug surfaced alongside it (Palm/
Passion Sunday wrongly reading true via Temporal_ef.named's own table
membership, then Christ the King wrongly reading false from an
over-broad fix) is closed with an explicit two-slug exclusion.

Domain-wide 1583-9999: every violet or Rose day is gloria=false except
the Easter Vigil (RG 431(c) lex specialis), every Requiem is
gloria=false, both measured exhaustively, zero exceptions. Mutation-
proved: disabling 431(c) reddens 8 tests including all four oracle
comparisons; disabling 238(c)'s feria-I-classis exclusion reddens
exactly the dedicated Ash Wednesday unit test, a genuine blind spot in
both oracle layers, reported rather than hidden.

Two open, cited findings, neither fixed here (out of this task's
"follow creed's exact seam" scope): a privileged Lenten/Passiontide
feria carrying one commemoration reads Gloria=true in the LMS Ordo but
Te-Deum=true/Gloria=false in FIUV -- the two oracles disagree with each
other, not merely with colitur (data/ef/expected-divergences-lms.sexp
L5, expected-divergences-fiuv.sexp F3); and a pre-existing, uncited
Colour.Violet bug on Rogation Monday/Tuesday in Temporal_ef.temporal,
surfaced by this comparison but root-caused as a separate defect (L6).

day/readings verified byte-identical against a build from the branch
tip before this task (v0.10.1's own tag predates an already-landed
bissextile fix that legitimately changed both, so it is not the right
baseline).

671 tests green (dune test); 678 with the exhaustive sweep
(COLITUR_EXHAUSTIVE_SWEEP=1 dune test --force, ~104s).

1 files changed, 101 insertions, 7 deletions

diff --git a/data/ef/expected-divergences-fiuv.sexp b/data/ef/expected-divergences-fiuv.sexp
index 5506e46..6ed9a04 100644
--- a/data/ef/expected-divergences-fiuv.sexp
+++ b/data/ef/expected-divergences-fiuv.sexp
@@ -15,12 +15,106 @@
 ; backs, the SAME convention every other allow-list file in this project
 ; uses: not always "colitur".
 ;
-; Empty as of this task: the Creed comparison found no unexplained
+; The Creed comparison (Task 6/Witnesses task) found no unexplained
 ; divergence anywhere in the 400-day window (399 comparable days, Holy
 ; Saturday excluded per test_creed_coverage -- see that test's own
-; citation). Kept as a file, not deleted, on the same footing as
-; expected-divergences-lms.sexp's own empty-but-live state after L1's
-; closure: the machinery that would explain a FUTURE divergence needs
-; somewhere to live, and an empty allow-list with a live "every entry
-; must be used" check is a real invariant worth keeping, not a
-; formality.
+; citation) and still does not. Phase 2 (celebrant-rubrics-phase1,
+; 2026-08-22) extended this file to two more axes, Gloria and Te Deum --
+; F1 below (Gloria) and F3 (Te Deum). Both prose-only, the same L2/L5/L6
+; convention data/ef/expected-divergences-lms.sexp's own header explains:
+; the real enforcement is test_fiuv_ordo.ml's own predicates and count
+; assertions, not a machine-loaded [id] record.
+;
+; F2 and F4 (Te Deum: every vigil; the three September Ember days) were
+; found, then CLOSED within the same task: tools/extract_fiuv_ordo.ml's
+; own Te Deum parser recognised only ONE of the source's two negative
+; phrasings ("non dicitur Te Deum"), so "sine Te Deum" -- how the source
+; actually negates a Sunday's/vigil's/Ember-day's own Te Deum -- fell
+; through to a bare "Te Deum" substring match and was wrongly read
+; [true]. Fixed in the extractor (its own citation has the full account);
+; the fixture was re-extracted from the SAME pdftotext dump; 24 of 400
+; rows flipped true->false, and colitur's own answer already matched the
+; corrected value on every one of them -- no code change needed for F2/
+; F4 specifically. See test/fixtures/fiuv-ordo-2025-2026.sexp's own
+; "RE-EXTRACTED" note and lib/rites/rite_ef/rubrics_ef.ml's own [te_deum]
+; header (237(a)'s comment) for the fuller account, including a real,
+; separate regression this same fix round found and reverted (a first
+; attempt at "every Sunday says Te Deum" wrongly excluded Christ the
+; King, caught by the LMS Ordo's own Gloria axis, 2024-10-27).
+
+; F1 -- OPEN, adjudicated FOR colitur. Fired once: 2026-04-03, Good
+; Friday.
+;
+; The raw source (docs/research/ordo/fiuv-ordo-2025-2026.pdf, page 46,
+; verified directly against the PDF, not merely the extracted fixture):
+; "3. Nig in Actione liturgica usque ad 4am partem, Viol. in S.
+; Communione. FERIA VI IN PASSIONE ET MORTE DOMINI, De ea, I cl. ...
+; Missa pr., (omittuntur ps. Iudica me et Gloria Patri), Gloria, sine
+; Credo, praef. comm." -- read LITERALLY, this states the Gloria IS said
+; on Good Friday. colitur says [false] (RG 432(d)'s own {!Colour.Black}
+; proxy this Mass shares with the Requiem population, {!Rite_ef.Rubrics_
+; ef.gloria}'s own header).
+;
+; ADJUDICATED FOR COLITUR: Good Friday's own liturgical action has no
+; Mass at all in the 1955-restored Holy Week (RG 28's own "non sit dies
+; liturgicus" framing for the Paschal Vigil is the closest parallel this
+; project already cites, {!Rite_ef.Rubrics_ef.creed}'s own RG 28-34
+; comment) -- no Consecration, no Communion under both species by the
+; celebrant, only Communion from the reserved Sacrament. That the Gloria
+; is NOT said on Good Friday is among the most widely and consistently
+; attested facts in the whole of 1962-rite literature (O'Connell,
+; Fortescue, every published Ordo this project has touched, and this
+; author's own independent knowledge of the rite) -- essentially
+; undisputed. The printed "Missa pr., ..., Gloria, ..." block, structured
+; identically to a REAL Mass's own rubric line elsewhere in this same
+; Ordo (compare Easter Day's "Missa pr., Gloria, sequentia, Credo, praef.
+; Pasch."), most plausibly describes the liturgical action's own
+; STRUCTURE using the book's standard notation for bookkeeping
+; consistency, or is a genuine single-word compiler error -- neither
+; resolvable without contacting the compiler, out of this task's own
+; scope. NOT the same failure mode as the Te Deum extractor bug above
+; (F2/F4): this is the RAW pdftotext TEXT itself making the claim, not a
+; downstream parsing artefact of test_fiuv_ordo.ml's own extraction --
+; re-run directly against a fresh -layout dump for this task, not merely
+; trusted from the fixture. The PDF's own page IMAGE was not separately
+; inspected (out of this task's own tool access), so a pdftotext-layer
+; misread (a stray line from an adjacent block folded into this one) is
+; not fully ruled out either -- named as a real possibility, not
+; silently excluded.
+
+; F3 -- OPEN, single-witnessed, colitur's own possible gap (Te Deum
+; axis). Fires 6 times, every one of the six dates
+; data/ef/expected-divergences-lms.sexp's own L5 already found for
+; Gloria: 2026-03-06/07/09/12/21/24, a privileged Lenten or Passiontide
+; feria (Class3, RG25) whose own office is impeded and carries exactly
+; one commemoration of a Class3 saint. colitur's [te_deum] reads [false]
+; (a plain ferial office by every branch {!Rite_ef.Rubrics_ef.te_deum}
+; checks); FIUV reads [true] in all six -- confirmed directly against
+; the raw pdftotext dump, not merely the fixture: e.g. 12 March, "Ll. 1a
+; et 2a (= 2a+3a) de Scr. occ., 3a de festo, Te Deum." -- a genuine,
+; unnegated "Te Deum" (neither of the two negative phrasings the
+; extractor now checks for), so this is NOT the same failure mode F2/F4
+; were.
+;
+; CITATION: the same n.302(b)/RG431(b) text L5 already cites (LT.txt --
+; a MISSAL rubric, not a Breviary one, so its literal text governs the
+; MASS's own Gloria, not Matins' Te Deum) does not directly license this
+; finding at all -- 237/238 (the BREVIARY rubric {!Rite_ef.Rubrics_ef.
+; te_deum} actually implements) has no clause naming a commemoration as
+; grounds for Te Deum. What this finding suggests, not yet established:
+; the DIVINE OFFICE may treat a commemorated day more generously than the
+; MASS does for exactly this shape (a commemorated saint's own lessons
+; are read at Matins even on an impeded day, which could plausibly extend
+; to Te Deum by an unwritten or unlocated convention) -- genuinely
+; speculative, offered as a research lead for a follow-on task, not a
+; citation this entry treats as settled.
+;
+; WHY NOT FIXED HERE: SINGLE-WITNESSED (the LMS fixtures never captured
+; Te Deum at all, so there is no second Ordo to corroborate this shape
+; the way the Sacred-Triduum/BVM-Saturday/vigil findings elsewhere in
+; this task were cross-checked). Fixing it would need the SAME kernel
+; signature widening L5 already declines for the identical reason
+; (`~commemorations` on {!Colitur_kernel.Rite.t.gloria}, and by extension
+; a parallel widening of [te_deum] if the two were ever to diverge in
+; their own inputs) -- out of a task briefed to follow creed's existing
+; seam exactly. Left OPEN for the same follow-on task L5 recommends.