feat(ef): implement RG 33's third omission trigger

RG 33 omits a II/III-class vigil in three cases: it falls on a Sunday, it
falls on a I-class feast, "vel si festum cui praemittitur in alium diem
transferri aut ad commemorationem reduci contingat". Only the first two
were built; the third was recorded in precedence_ef.ml as unimplemented on
the grounds that no witness existed in the shipped data. That reasoning was
wrong, and the rule fires on 1 744 days across 1583-9999.

Both halves of the clause reduce to one observable question -- is the feast
the OBSERVED office on the following day (RG 34 puts it there) -- so the
kernel asks it once per candidate, after place_transfers has settled the
year. No fixed point is needed: a vigil is a candidate only on its own day,
never on its feast's, so suppressing it cannot change what the next day
observes. Precedence.rules gains vigil_feast, which returns the slug of the
feast a vigil precedes; the kernel cannot infer that itself, because only
two of the five vigil/feast pairs share a slug stem.

Blast radius, measured pre-change binary vs HEAD over the whole domain and
classified: 1 744 days, three shapes, zero unexplained. 1 199 are the feast
reduced to a commemoration (10 August on a Sunday, St Lawrence); 478 and 67
are the feast transferred under RG 96 after the Sacred Heart or Corpus
Christi takes its day. The Assumption's and the Ascension's vigils never
qualify -- their I-class feasts always keep their own day.

Independently witnessed, which is unusual here. The published Ordo -- the
only witness outside the Divinum Officium -> missalemeum -> lectio lineage
-- omits St Lawrence's vigil on 2025-08-09, agreeing with colitur against
both engines. That date had been read earlier as an Ordo gap; the Ordo was
right, and correcting the misreading is what surfaced this clause. On
2027-08-09 the feast does keep its day and the Ordo omits a vigil colitur
correctly keeps, which is a genuine Ordo gap.

Allow-lists: C39 (lectio, 10 rows) and a 2038 oracle class citing the
register, the 2026-2027 window having no instance. The golden pin asserting
St Lawrence's vigil is violet moved 2025 -> 2027; its own comment had
reasoned about the vigil's weekday and missed that RG 33 also looks at the
feast's. Two new pins cover both shapes of the clause.

The vigil/feast table is built with Slug.of_string_exn: mutation testing
showed that of_string plus Result.to_option turns a typo into None, which
this hook's contract reads as "not a vigil", switching the rule off in
silence. Two tests assert the table against the shipped data in both
directions.

1 files changed, 72 insertions, 8 deletions

diff --git a/test/test_golden.ml b/test/test_golden.ml
index 33ca362..b882596 100644
--- a/test/test_golden.ml
+++ b/test/test_golden.ml
@@ -1327,14 +1327,70 @@ let test_rogation_wednesday_yields_to_a_feast_2026 () =
    white" -- which colitur already emits correctly (ef-ascension-vigil, white,
    pinned by the two Rogation Wednesday cases just above).
 
-   Pinned in years where each vigil is actually OBSERVED rather than impeded:
-   9 August 2025 is a Saturday and 14 August 2025 a Thursday, both checked
-   with `date -d`. In 2026, 9 August is a Sunday and the vigil never appears. *)
-let test_vigil_of_st_lawrence_is_violet_2025 () =
-  check ~msg:"RG 128: III-class vigil outside Paschaltide is violet, not red" 2025 8 9
-    "2025-08-09 saturday season=time-after-pentecost week=8 slug=vigil-of-st-lawrence \
+   Pinned in years where each vigil is actually OBSERVED rather than impeded.
+   14 August 2025 is a Thursday and the Assumption always keeps 15 August (I
+   class, nothing in the universal calendar outranks it), so that pin is
+   stable in any year the vigil's own day is free.
+
+   ST LAWRENCE'S PIN MOVED 2025 -> 2027 when RG 33's third omission trigger
+   was implemented. Its original comment reasoned only about the vigil's OWN
+   weekday -- "9 August 2025 is a Saturday", true, so RG 33's Sunday trigger
+   does not fire -- and missed that the rule also looks at the FEAST: 10
+   August 2025 is a Sunday, the II-class Sunday reduces St Lawrence to a
+   commemoration, and "vel si festum cui praemittitur... ad commemorationem
+   reduci contingat" omits the vigil entirely. The day now keeps the BVM
+   Saturday Office instead, pinned separately just below. 9 August 2027 is a
+   Monday and 10 August 2027 a Tuesday (both `date -d`-checked), so St
+   Lawrence keeps his own day there and the vigil genuinely is observed --
+   which is what this pin has always been trying to assert about RG 128's
+   colour rule. *)
+let test_vigil_of_st_lawrence_is_violet_2027 () =
+  check ~msg:"RG 128: III-class vigil outside Paschaltide is violet, not red" 2027 8 9
+    "2027-08-09 monday season=time-after-pentecost week=12 slug=vigil-of-st-lawrence \
      rank=class-3 colour=violet subject=saint name_la=- comms=[romanus:ordinary] in=- out=[]"
 
+(* RG 33's THIRD omission trigger -- "vel si festum cui praemittitur in alium
+   diem transferri aut ad commemorationem reduci contingat" -- on the shipped
+   universal calendar, no overlay involved. 10 August 2025 is a Sunday; the
+   II-class Sunday outranks St Lawrence's II-class feast (RG 15), so the feast
+   is "ad commemorationem reducta" on its own day and the vigil of 9 August is
+   "penitus omissa".
+
+   This is the clause's most common live shape by a wide margin: 1 199 of the
+   1 744 days it fires on across 1583-9999 are St Lawrence's vigil dropped
+   because 10 August fell on a Sunday. The day does not become a bare feria --
+   9 August 2025 is a Saturday, so RG 78's votive Office of the BVM takes the
+   freed IV-class Saturday, which is why this pin reads white and subject=bvm.
+   St Romanus keeps his commemoration either way (RG 111).
+
+   Independently witnessed. The published Ordo -- the project's only witness
+   outside the Divinum Officium -> missalemeum -> lectio lineage -- omits the
+   vigil on this exact date, and lectio (and so missalemeum) keeps it; both
+   allow-lists gained cited entries for that disagreement. *)
+let test_rg33_vigil_omitted_when_its_feast_is_commemorated_2025 () =
+  check
+    ~msg:"RG 33: 10 Aug 2025 is a Sunday, so St Lawrence is only commemorated and his vigil is \
+          omitted entirely -- the BVM Saturday Office (RG 78) takes the freed day"
+    2025 8 9
+    "2025-08-09 saturday season=time-after-pentecost week=8 \
+     slug=ef-time-after-pentecost-8-saturday rank=class-4 colour=white subject=bvm \
+     name_la=Officium sanctae Mariae in sabbato comms=[romanus:ordinary] in=- out=[]"
+
+(* The same clause's OTHER half, "in alium diem transferri". 24 June 2038 is
+   impeded by the Sacred Heart (I class, movable), so the Nativity of St John
+   the Baptist is TRANSFERRED under RG 96 and its vigil of 23 June is omitted.
+   478 days across the domain take this shape (a further 67 via Corpus
+   Christi). Pinned in 2038 because that year already carries an oracle
+   fixture, so the divergence this creates is visible to layer 4 as well. *)
+let test_rg33_vigil_omitted_when_its_feast_transfers_2038 () =
+  check
+    ~msg:"RG 33: the Sacred Heart takes 24 June 2038, St John the Baptist transfers, \
+          and his vigil of 23 June is omitted entirely"
+    2038 6 23
+    "2038-06-23 wednesday season=time-after-pentecost week=1 \
+     slug=ef-time-after-pentecost-1-wednesday rank=class-4 colour=green subject=temporal \
+     name_la=- comms=[] in=- out=[]"
+
 let test_vigil_of_the_assumption_is_violet_2025 () =
   check ~msg:"RG 128: II-class vigil outside Paschaltide is violet, not white" 2025 8 14
     "2025-08-14 thursday season=time-after-pentecost week=9 slug=vigil-of-the-assumption \
@@ -1342,8 +1398,16 @@ let test_vigil_of_the_assumption_is_violet_2025 () =
 
 let suite =
   ( "golden pins (known-tricky years)",
-    [ Alcotest.test_case "RG128: St Lawrence's vigil is violet (2025)" `Quick
-        test_vigil_of_st_lawrence_is_violet_2025;
+    [ Alcotest.test_case "RG128: St Lawrence's vigil is violet (2027)" `Quick
+        test_vigil_of_st_lawrence_is_violet_2027;
+      Alcotest.test_case
+        "RG33 third trigger: the vigil is omitted when its feast is reduced to a commemoration \
+         (2025-08-09, St Lawrence)"
+        `Quick test_rg33_vigil_omitted_when_its_feast_is_commemorated_2025;
+      Alcotest.test_case
+        "RG33 third trigger: the vigil is omitted when its feast is transferred away \
+         (2038-06-23, St John the Baptist)"
+        `Quick test_rg33_vigil_omitted_when_its_feast_transfers_2038;
       Alcotest.test_case "RG128: the Assumption's vigil is violet (2025)" `Quick
         test_vigil_of_the_assumption_is_violet_2025;
       Alcotest.test_case "Easter extreme: 1598 earliest (22 Mar, Gauss-verified)" `Quick