fix(ordo): correct L5/F3 misadjudication -- Cum Sanctissima, not a gap

data/ef/expected-divergences-lms.sexp's L5 (16 dates, LMS Ordo Gloria=true
vs colitur's false) and data/ef/expected-divergences-fiuv.sexp's F3 (6
dates, FIUV Te Deum=true vs colitur's false) were both misadjudicated as
colitur's own gap under an RG 302(b)/431(b) hypothesis. Both are wrong.

These are occurrence differences from the 2020 decree Cum Sanctissima,
which colitur excludes from its 1962 core by binding design (CLAUDE.md
decision 2). RG 303(b) ("dici potest tantum si occurrit dies liturgicus IV
classis") already prohibits 302(b) on the Class3 privileged ferias both
entries cover -- the original text quoted 303(b) but never followed it to
its conclusion. Decisively, the FIUV Ordo prints BOTH readings on the six
shared dates, separated by a literal "-VEL-" token: the primary, strict-
1962 reading (feria wins, sine Gloria, no Te Deum) matches colitur exactly;
the second, Cum-Sanctissima reading (saint kept, Gloria, Te Deum) matches
the LMS Ordo exactly.

F3 itself was never a real Te Deum finding: tools/extract_fiuv_ordo.ml's
extract_te_deum scanned each block unbounded for "Ad Mat.", so on these six
days -- whose primary office states no "Ad Mat." of its own -- it silently
read the second office's own Te Deum instead. Fixed with a new
split_at_alt_office, cutting every block at "-VEL-" before any field
extraction runs; the alternative is captured, not discarded, in a new
alt_te_deum field. A hard characterisation probe asserts the split fires on
exactly the six known dates.

Auditing for the same trap elsewhere (this is the third such manufactured
divergence in this project) found a related but narrower marker, "Vel (sec.
decretum <<Cum sanct.>>) Missa ...", offering an alternate Mass formulary
within a single office. It never corrupted Gloria/Credo (always stated
before the marker on all 20 real instances, confirmed by diff), only the
uncompared praef field; hardened defensively anyway, before it could ever
fire live. The LMS Ordo (all three editions) carries no equivalent
dual-print mechanism at all. expected-divergences-missalemeum.sexp's M27
is the same genus but was already correctly adjudicated, not touched.

Both allow-list entries are rewritten, not deleted: the original SHAPE/
CITATION/WHY-NOT-FIXED-HERE reasoning is preserved verbatim as a recorded,
superseded hypothesis, matching this project's own correction convention.
test_lms_ordo.ml's matching predicate is unchanged (the 16-date population
still fires, and always will, because it reflects a standing exclusion,
not a bug); only its citation comment is corrected. test_fiuv_ordo.ml's
F3-specific predicate/count assertion is removed, matching this file's own
prior F2/F4 closure discipline.

lib/ is untouched. dune test: 676/676. make check
(COLITUR_EXHAUSTIVE_SWEEP=1, full 1583-9999 domain): 678/678.

1 files changed, 122 insertions, 18 deletions

diff --git a/data/ef/expected-divergences-lms.sexp b/data/ef/expected-divergences-lms.sexp
index df813f9..d91028a 100644
--- a/data/ef/expected-divergences-lms.sexp
+++ b/data/ef/expected-divergences-lms.sexp
@@ -10,6 +10,11 @@
 ; Gloria (RG431-432) -- both fire in more than one window, at DIFFERENT
 ; counts per window, so both are prose-only (L2's own precedent), not
 ; machine-checked [id] records; see each for the real enforcement site.
+; L5 was MISADJUDICATED as "colitur's own gap" and CORRECTED
+; (l5-f3-readjudication task, 2026-08-22): see L5's own entry for the
+; full account -- it is an OCCURRENCE difference (the LMS Ordo follows
+; the 2020 decree Cum Sanctissima, colitur excludes it by binding
+; design), not a Gloria defect.
 ;
 ; A SEPARATE file from data/ef/expected-divergences.sexp (the lectio
 ; allow-list) and data/ef/expected-divergences-missalemeum.sexp (the
@@ -233,14 +238,32 @@
  (note "2023-12-24 (the 4th Sunday of Advent AND Christmas Eve, a genuinely rare coincidence -- roughly 1 year in 7 -- absent from both the 2024-2025 and 2025-2026 windows, where Christmas Eve falls on a Tuesday and a Wednesday respectively). colitur says the Creed IS said (true), reading [temporal]'s own weekday (Sunday) per RG475(a)'s clause, positioned ahead of the vigil-exclusion check in {!Rite_ef.Rubrics_ef.creed} -- unconditionally, the same way it already does for RG16(a)'s Feast-of-the-Lord case. The Ordo says \"No Gl No Cr\" for this date. ADJUDICATED FOR COLITUR, not merely left open by default: RG30(a)'s own \"locum tenet ... nulla fit commemoratio\" language is textually closer to RG16(a) (which colitur already treats as Creed-preserving) than to an ordinary vigil superseding a Sunday -- the Vigil does not merely WIN THE DAY, it takes the SUNDAY'S OWN IDENTITY, the same shape RG475(a) was written to reach. The COUNTER-reading is real and not dismissed here: RG475(a)'s own text says \"locum cedat ... festo\" (gives way to a FEAST), and RG35 (Caput VI) classifies feria/vigilia/festum as three distinct liturgical-day categories -- a strict textual reading could confine 475(a)'s override to FEAST-displacement only, leaving a VIGIL-displaced Sunday to fall through to the general vigil exclusion (476, via RG33's own \"a vigil is a distinct category\" reasoning already coded elsewhere in {!Rite_ef.Rubrics_ef.creed}). Left OPEN rather than closed on this task's own authority: adjudicating scripture-grade primary-source questions is out of scope for a witness-wiring task, and the two readings are both defensible. Not fixable inside this task's own no-lib-changes constraint regardless of which reading eventually wins.")
  (expected_rows 1))
 
-; L5 -- OPEN, colitur's own gap, found extending this suite to the Gloria
-; axis (celebrant-rubrics-phase1 Phase 2, 2026-08-22). NOT an active
-; machine-checked [id] entry -- same reasoning L2 above gives: the count
-; varies window to window (5 in 2023-2024, 5 in 2024-2025, 6 in
-; 2025-2026), so a single static [expected_rows] cannot check it. The REAL
-; enforcement is test_lms_ordo.ml's own [is_l5_lenten_commemoration]
+; L5 -- MISADJUDICATED, NOW CORRECTED (l5-f3-readjudication task,
+; 2026-08-22). This entry previously read "OPEN, colitur's own gap" --
+; WRONG, on two independent grounds (below). It is NOT a Gloria defect
+; of colitur's. It is NOT a rubric dispute at all: it is an OCCURRENCE
+; difference, because the LMS Ordo is compiled to the 2020 decree Cum
+; Sanctissima (its own introduction says so, docs/research/ordo/
+; PROVENANCE-lms.md) and colitur excludes that decree from its 1962 core
+; by BINDING DESIGN (CLAUDE.md decision 2: "Every addition [2020 Quo
+; Magis/Cum Sanctissima, any community's proper] is an overlay, never
+; core"). On these 16 dates colitur and the LMS Ordo each correctly
+; describe a DIFFERENT, both genuine, calendar -- they were never
+; disagreeing about the same one. VERDICT: colitur correct. NOT an
+; active machine-checked [id] entry -- same reasoning L2 above gives:
+; the count varies window to window (5 in 2023-2024, 5 in 2024-2025, 6
+; in 2025-2026), so a single static [expected_rows] cannot check it. The
+; REAL enforcement is test_lms_ordo.ml's own [is_l5_lenten_commemoration]
 ; predicate plus [make_suite]'s own ~expected_gloria_l5 parameter,
-; asserted per window directly.
+; asserted per window directly -- KEPT, unlike a bug-driven divergence
+; that stops firing once fixed (data/ef/expected-divergences-fiuv.sexp's
+; own F3, closed by this same task, is that other shape): this one is
+; EXPECTED TO RECUR IN EVERY FUTURE WINDOW, because the exclusion it
+; reflects is a standing design decision, not a defect to remove.
+;
+; THE ORIGINAL, SUPERSEDED HYPOTHESIS -- preserved below verbatim, the
+; same "retain the wrong reading and say why it was wrong" convention
+; L3's and F1's own history in this file already establish, NOT deleted:
 ;
 ; SHAPE, identical in all 16 instances found (5+5+6, zero exceptions):
 ; a privileged Lenten or Passiontide feria (Class3, violet -- RG25,
@@ -251,6 +274,8 @@
 ; day is a plain ferial office by every test [Rite_ef.Rubrics_ef.te_deum]
 ; runs -- no Sunday, no vigil, no octave, [named] = [None], [subject] =
 ; [Temporal]); the Ordo reads [true] in every one of the 16 instances.
+; (This description of the SHAPE is still accurate -- what was wrong is
+; the CONCLUSION drawn from it, below.)
 ;
 ; CITATION (docs/research/LT.txt:2359-2361, "V - De Missis festivis",
 ; quoted in rubrics_ef.ml's own [gloria] header in full): "302. Sensu
@@ -263,7 +288,9 @@
 ; ferial Office, carrying a commemoration, is elevated to "festive" status
 ; for the Gloria specifically.
 ;
-; WHY NOT FIXED HERE: n.303(b), the very next paragraph (LT.txt:2359-2367,
+; WHY NOT FIXED HERE (the original text, which already quoted the very
+; clause that defeats its own hypothesis, below, but did not follow it
+; to its conclusion): n.303(b), the very next paragraph (LT.txt:2359-2367,
 ; same quote in rubrics_ef.ml), qualifies n.302(b)'s own Masses: "dici
 ; potest tantum si occurrit dies liturgicus IV classis" -- CAN ONLY BE
 ; SAID if the liturgical day occurring is IV class. Every instance found
@@ -273,29 +300,29 @@
 ; project's sources), which may clarify that 301-303 govern which Mass a
 ; priest may freely CHOOSE among several options in one church on one day
 ; (a dimension colitur does not model -- see [gloria]'s own SCOPE NOTE),
-; a different question from \"does THE Mass actually said, once
-; determined, say the Gloria\" (431(b)'s own question). Implementing this
+; a different question from "does THE Mass actually said, once
+; determined, say the Gloria" (431(b)'s own question). Implementing this
 ; also has a real architectural cost this task's own instructions
 ; explicitly avoided: [Rite_ef.Rubrics_ef.gloria] would need a FOURTH
 ; parameter, the day's own {!Colitur_kernel.Liturgical_day.t.commemorations}
 ; -- a kernel [Rite.t.gloria] signature change, the same shape RG16(a)'s
 ; own [admit] fix needed a [~temporal] parameter for, and squarely out of
-; a task briefed to \"follow exactly how creed was wired -- same seam\".
+; a task briefed to "follow exactly how creed was wired -- same seam".
 ; Left OPEN, not guessed into a possibly-wrong rule. Recommended for a
 ; follow-on task, ideally alongside Phase 3's Preface work (RG483 governs
 ; a DIFFERENT, but textually adjacent, commemoration-vs-Mass-part
 ; question, so the two may share the same kernel signature widening).
 ;
-; UPDATE (same task, after wiring the FIUV Ordo's own Gloria and Te Deum
-; axes, test_fiuv_ordo.ml): this paragraph previously said FIUV's own
-; [gloria] axis was not compared at all. It now is, and the result
+; UPDATE (same original task, after wiring the FIUV Ordo's own Gloria and
+; Te Deum axes, test_fiuv_ordo.ml): this paragraph previously said FIUV's
+; own [gloria] axis was not compared at all. It now is, and the result
 ; WEAKENS this entry's own case rather than strengthening it: FIUV's own
 ; [gloria] is [false] on all six of these exact dates in its window --
 ; AGREEING WITH COLITUR, DIRECTLY CONTRADICTING the LMS Ordo's [true]
-; above. So this is not \"colitur has a corroborated gap\", it is \"two
+; above. So this is not "colitur has a corroborated gap", it is "two
 ; independent Ordos disagree with EACH OTHER on Gloria for this shape,
 ; and colitur happens to agree with the universal one, not the diocesan
-; one\". What FIUV's OWN Te Deum axis DOES show for the identical six
+; one". What FIUV's OWN Te Deum axis DOES show for the identical six
 ; dates: [true] -- Te Deum, unlike Gloria, genuinely IS said at Matins on
 ; these days (data/ef/expected-divergences-fiuv.sexp's own F3), a real,
 ; single-witness finding that this entry's own n.302(b)/431(b) citation
@@ -303,8 +330,85 @@
 ; mean L5 itself is answering the wrong question, not merely open.
 ; Genuinely uncertain which of the two Mass-Gloria witnesses (LMS or
 ; FIUV) is right for this shape; left OPEN, not adjudicated either way,
-; for a follow-on task with more primary-source access than this one
-; had.
+; for a follow-on task with more primary-source access than this one had.
+; (CORRECTED, l5-f3-readjudication task: F3's own "single-witness
+; finding" was never a Te Deum finding at all -- it was
+; tools/extract_fiuv_ordo.ml reading the WRONG PRINTED OFFICE, see below.
+; There was no Office-vs-Mass question to be uncertain about.)
+;
+; THE CORRECTION, in full (l5-f3-readjudication task, 2026-08-22):
+;
+; 1. RG303(b) ALREADY DEFEATS the n.302(b)/431(b) hypothesis on its own
+;    terms -- the original text above quoted it but stopped short of the
+;    conclusion. Quoted again, in full, because it is the load-bearing
+;    sentence: "dici potest tantum si occurrit dies liturgicus IV
+;    classis" -- [a Mass "of a commemoration occurring in the day's own
+;    Office"] MAY BE SAID ONLY IF the liturgical day occurring is FOURTH
+;    class. A privileged Lenten or Passiontide feria is THIRD class
+;    (RG25) -- every one of these 16 instances. So n.302(b) is
+;    PROHIBITED, not merely inapplicable, on precisely the days this
+;    entry covers, and 431(b)'s own Gloria elevation, which depends on
+;    302(b), can never reach them. Not "unresolved without n.304" as the
+;    original text guessed: 303(b) alone already closes the question,
+;    the wrong way for the original hypothesis.
+;
+; 2. THE FIUV ORDO PRINTS BOTH OPTIONS, and this is decisive on its own,
+;    independent of (1). All 16 LMS dates are Lenten/Passiontide
+;    privileged ferias; six of them fall inside the FIUV fixture's own
+;    2025-2026 window (2026-03-06/07/09/12/21/24 -- the same six F3
+;    named). For every one of the six, the raw FIUV source (verified
+;    directly against a fresh pdftotext -layout dump, not merely the
+;    fixture) prints TWO complete readings for the SAME civil date,
+;    separated by a literal "-VEL-" ("or") token. 24 March 2026, in
+;    full:
+;
+;      24. Viol  Feria III post Dom. I Passionis, De ea, III cl. - Off. feriale temp. Passionis.
+;                Ad LAUDES ... comm. S Gabrielis Archangeli.
+;                Missa pr., (omittuntur ps. Iudica me et Gloria Patri), sine Gloria, comm. S
+;                Gabrielis Archangeli, sine Credo, praef. de Sancta Cruce or. super populum.
+;                I VESPERAE de seq., comm. huius diei. - Compl. de dom.
+;                -VEL-
+;                Feria III. S Gabrielis Archangeli, III cl. (Priv.) - Off. ordinarium. - Ad Mat.
+;                Ll. 1a et 2a (= 2a + 3a) de Scr. occ., 3a de festo, Te Deum. - Ad LAUDES comm.
+;                Feriae.
+;                Missa pr., Gloria, comm. Feriae, tractus, sine Credo, praef. de Sancta Cruce vel
+;                de Angelis.
+;
+;    The FIRST option is strict 1962: the feria wins, "sine Gloria", no
+;    Te Deum stated, the saint merely commemorated -- EXACTLY colitur's
+;    own answer, field for field. The SECOND, after "-VEL-", is the
+;    Cum-Sanctissima-elevated option: the saint's own feast, "Off.
+;    ordinarium", Gloria, Te Deum, the feria commemorated instead --
+;    EXACTLY the LMS Ordo's own answer. The FIUV Ordo's own title page
+;    states the convention this book uses throughout for exactly this
+;    distinction (N.B., page 1, verified directly against the PDF):
+;    "days of the third class to which the norms of the decree Cum
+;    santissima [sic] apply are marked 'III cl.' ... but days ... to
+;    which these norms do not apply are marked 'III cl. (Priv.)'." One
+;    document, compiled by a THIRD party independent of both colitur's
+;    own lineage and the LMS, prints BOTH engines' own answers side by
+;    side as two legitimate readings of two different calendars. There
+;    is no dispute left to adjudicate.
+;
+; So the Gloria and Te Deum RULES were never in dispute at all: each
+; engine's answer follows correctly from its own occurrence result, on
+; its own calendar. What this entry originally described as "colitur's
+; own gap" is colitur correctly implementing the 1962 core it was built
+; to implement, diverging from a printed book that, by its own
+; introduction, deliberately implements something else.
+;
+; NOTE ALSO (so a reader arriving here from
+; data/ef/expected-divergences-fiuv.sexp's own F3 does not draw the
+; wrong conclusion from that entry's own closure): F3 did NOT
+; "corroborate" or "contradict" L5 at any point in this history. Both
+; were the same underlying fact read two different ways, and F3's own
+; apparent Te Deum-true reading was never a real second data point in
+; the first place -- it was tools/extract_fiuv_ordo.ml's own
+; [extract_te_deum] silently reading the SECOND ("-VEL-") office's own
+; Te Deum instead of the primary's, on exactly these six dates, closed
+; by fixing the extractor (tools/extract_fiuv_ordo.ml's own
+; [split_at_alt_office]), not by resolving any Office-vs-Mass question.
+; See F3's own entry for the full account.
 
 ; L6 -- OPEN, root-caused to a DIFFERENT, PRE-EXISTING bug this Gloria
 ; comparison merely surfaced, not a Gloria defect itself. Fired once,