test(differential): fix round 1 -- Holy Week, Lent Ember, RG 69

Two Critical findings, both real: Holy Week (including the entire Sacred
Triduum) was reading Passion Week's Masses -- the original colitur_keys
widened lectio's single, conflated ini section into both weeks instead of
Passion week alone. Narrowed to Passion week only; Holy Week's own four
cleanest days (Monday, Tuesday, Thursday, Saturday) are hand-authored
directly from the Missal, verified twice against both scans (C23); Holy
Wednesday and Good Friday are genuinely multi-lesson liturgies with no
single Epistle slot and are deliberately left unanswered, not guessed (C24).

The Lenten Ember days were still dead keys: Temporal_ef computes
ef-lent-ember-{wed,fri,sat}, lectio's ini names the same data
ef-lent-1-{wednesday,friday,saturday}, and colitur_keys never translated
it -- both engines independently fell through to the same wrong ferial
answer, so even the differential stayed green (C22, a genuine lectio bug
recorded for upstream). Fixed the instance and the class: colitur_keys
gained the rename, and tools/bootstrap_lectionary.ml gained
assert_reachable, sweeping Rite_ef.Temporal_ef.temporal at generation time
so a translation table entry naming a slug colitur never computes dies
loudly instead of shipping silently, plus an informational log of every
real slug with no entry at all.

C21 ("colitur's step 3 cannot reach a correct answer for the fixed
Christmastide slugs") is retired, superseded rather than merely patched:
the Nativity-Octave days now have their own direct Missal formulary
(Tit. 3,4-7/Luc. 2,15-20, "Diebus infra octavam Nativitatis Domini"),
resolved at step 2, not step 3 at all. That surfaced a real RG 69 gap
underneath it: a Sunday landing 26-31 December should keep its own Office,
not the fixed weekday placeholder, which Temporal_ef does not currently
distinguish. C6's verdict is re-opened (open, not colitur) for the 19 of
its 138 rows where this applies directly, and a new entry C25 (57 rows)
records the same gap propagating one hop further through step 3 into Holy
Name week. Both are out of this task's own scope -- a behaviour change to
a shared, multi-round-reviewed kernel-adjacent function -- and recorded,
not fixed, per the coordinator's own instruction.

C6's own note is rewritten in full rather than layered with another
correction; C18's stale trailing count is fixed to match its own pin; the
Vigil-misapplication risk flagged as a Minor closed itself once the
Nativity-Octave fix landed, confirmed by inspection rather than assumed.

dune test: 358/358. COLITUR_EXHAUSTIVE_SWEEP=1 dune test --force: 359/359.

1 files changed, 127 insertions, 51 deletions

diff --git a/test/test_differential.ml b/test/test_differential.ml
index 74cc54d..bdb3377 100644
--- a/test/test_differential.ml
+++ b/test/test_differential.ml
@@ -717,54 +717,117 @@ let lectio_no_reading_slugs =
   [ "vincent-ferrer"; "francis-of-paola"; "isidore-of-seville"; "thomas-aquinas"; "john-of-god";
     "sts-felicitas-perpetua"; "frances-rome" ]
 
-(* C21 (task 8) -- [verdict open], the FIRST entry in this file that is not
-   [verdict colitur] (see data/ef/expected-divergences.sexp's own header for
-   the file-level rule this departs from, and the note on why). A genuine,
-   precisely-diagnosed colitur LOGIC LIMITATION, not a data/bootstrap issue
-   and not a rubric dispute: {!Rite_ef.Lectionary_ef.readings}'s step 3 is a
-   SINGLE hop to the immediately preceding civil Sunday's own temporal
-   identity. For a run of colitur-only FIXED Christmastide slugs
-   ([ef-nativity-octave-day-{5,6,7}], 29-31 December; the
-   [ef-christmas-{0,1,2}-<weekday>] ferias of the two weeks that follow) that
-   ONE Sunday is itself frequently ANOTHER entry-less fixed slug (another
-   Nativity-Octave day, or, on the single occasion 24 December itself falls
-   on a Sunday, [ef-nativity-vigil]) -- so step 3 finds nothing (95 of these
-   107 rows) or, on the 12 rows where the walk happens to land exactly on
-   that one Vigil-is-a-Sunday collision, finds the WRONG thing (the Vigil's
-   own Mass, misapplied to a later feria that is not itself Christmas Eve).
-   lectio's own resolution reaches further by some mechanism this task did
-   not fully reverse-engineer (empirically: it variously finds Christmas
-   Day's own Mass or Advent IV's own Mass, depending on the year's
-   alignment). Properly fixing this needs either an iterative walk-back in
-   [readings] itself or a per-year-verified direct entry for every fixed
-   Christmastide slug (task 6's own rigor, applied to a new population) --
-   both are logic/behaviour changes to a shared, carefully-reviewed
-   kernel-adjacent function (Tasks 4-6's own multi-round review history),
-   out of this task's own safe scope. Recorded here, not left silently
-   failing, because the orchestrating instruction's own condition for
-   "leave it failing" is "if you cannot explain a divergence" -- this one
-   IS explained, precisely; what remains unresolved is which SAFE fix to
-   build, not what the mechanism is. Gated on the two literal colitur
-   slug-family prefixes (never a bare "any citation diff" rule) so an
-   unrelated future bug on any OTHER slug still surfaces as unexplained.
-   Derived directly from the OCaml comparator's own failure output
-   (test_layer_c_counts_match_citations), not hand-counted first and
-   cross-checked after: 107. NOTE: an earlier version of this comment (and
-   of [colitur_keys] in tools/bootstrap_lectionary.ml) measured 202 here,
-   before a regression in the FIRST version of the "ef-christmas-sunday-0"
-   translation (replaced instead of widened the key -- caught by
-   test_lectionary_ef.ml's own pre-existing test, not by this reasoning)
-   was fixed; fixing it also legitimately shrank this population by 95
-   rows. data/ef/expected-divergences.sexp's own C21/C6 notes have the full
-   before/after account, including the ONE guess this task tried for
-   [ef-nativity-octave-day-{5,6,7}] (Holy Name Sunday's own citation,
-   uniformly) and reverted once measured -- Missal-confirmed
-   (docs/research/scan1.txt, "Dominica infra octavam Nativitatis Domini,
-   II classis": Gal. 4,1-7 / Luc. 2,33-40) to be right for the genuine
-   Sunday-within-the-Octave case alone, wrong whenever the true preceding
-   Sunday is Christmas Day or Advent IV instead. *)
-let christmastide_fixed_slug_family (c : row) =
-  starts_with ~prefix:"ef-nativity-octave-day-" c.slug || starts_with ~prefix:"ef-christmas-" c.slug
+(* C21 -- RETIRED, task 8 fix round 1 (coordinator review). Previously a
+   [verdict open] entry (107 rows) for "colitur's step 3 cannot reach a
+   correct answer for the fixed Christmastide slugs". That diagnosis has
+   been SUPERSEDED, not merely fixed around: Important 3(a)'s own direct
+   Missal formulary (tools/bootstrap_lectionary.ml's own
+   [nativity_octave_entries]) gives [ef-nativity-octave-day-{5,6,7}] a
+   real, correct answer via STEP 2, not step 3 at all, so "step 3 cannot
+   reach it" is no longer even true. What C21 used to cover has split
+   cleanly into two DIFFERENT, more precisely diagnosed populations:
+   colitur's own [ef-nativity-octave-day-N] rows now fold into C6 above
+   (widened, [verdict] re-opened -- see its own note); the
+   [ef-christmas-1-<weekday>] rows that resume THAT slug's own (partly
+   wrong, per RG 69) answer one hop later via step 3 are C25 below, a
+   genuinely different id because the mechanism is now precisely RG 69
+   propagating through step 3, not "step 3 has no answer". No predicate
+   or helper of this file's own still names "C21" -- removing the code
+   alongside the id is the same discipline this file's own header already
+   applies to C2-C5/C7/C9-C13 (closed, cited, never left as dead code). *)
+
+(* C22 (task 8, fix round 1, coordinator review, Critical 2): the Lenten
+   Ember days (RG 91 entry 18, colitur's own ef-lent-ember-{wed,fri,sat})
+   are reachable for the first time this fix round -- tools/
+   bootstrap_lectionary.ml's own [colitur_keys] comment has the full
+   account of the dead-key bug this closes. lectio's own SLUG for these
+   three days already matched colitur's (both call it "ef-lent-ember-wed"
+   etc, confirmed by this row's own [c.slug]/[l.slug] agreement -- no
+   [Slug_f] in this entry's own diff set), but lectio's own READING
+   resolution has the IDENTICAL internal mismatch colitur's used to have:
+   its own ini keys this data under "ef-lent-1-<weekday>", and its own
+   [caldata.Readings] never reaches it either, so it still falls back to
+   Lent I Sunday's own Mass (e.g. "2 Cor. 6:1-10"/"Matt 4:1-11") on every
+   one of these dates. That is a genuine LECTIO bug, reported here for
+   upstream, NOT fixed in this task (the brief's own triage rule: record,
+   do not patch the sibling project) -- and it is exactly the "Holy
+   Thursday was violet in both" shape this project's own CLAUDE.md already
+   names: layer 3 was STRUCTURALLY BLIND to this the whole time colitur
+   also had the bug, because both engines independently produced the same
+   wrong answer. Gated on colitur's own slug family (a literal 3-member
+   list, not a pattern) and the diff set staying [First_f; Gospel_f]
+   alone (slug/rank/colour already agree). Derived directly from the OCaml
+   comparator's own failure output, not hand-counted first and cross-
+   checked after: 128 (43 + 43 + 42 -- a few of the 46-year population are
+   impeded by a competing sanctoral feast each year, so the Ember office
+   itself is not always the observed day). *)
+let lent_ember_slugs = [ "ef-lent-ember-wed"; "ef-lent-ember-fri"; "ef-lent-ember-sat" ]
+
+(* C23/C24 (task 8, fix round 1, coordinator review, Critical 1): Holy
+   Week's own citations (tools/bootstrap_lectionary.ml's own
+   [holy_week_entries] comment has the full Missal citations and the two-
+   scan corroboration for each). lectio has no Holy Week propers at all --
+   every "ef-passiontide-0-<weekday>" ini section is Passion week's own
+   Mass (or, for Tuesday specifically, Holy Tuesday's own Passion Gospel
+   riding along on Passion week's own Epistle -- a genuine internal
+   inconsistency in lectio's own source, not this project's to fix), so
+   lectio shows THAT reused/conflated citation on every Holy Week weekday,
+   every year, regardless of what the real Missal prints there.
+
+   C23 covers the four days colitur now answers correctly and completely
+   (Monday, Tuesday, Thursday, Saturday) -- verdict colitur, cited
+   directly. C24 covers the two days colitur deliberately answers with []
+   instead of a guess (Wednesday, Good Friday -- both genuinely
+   multi-lesson liturgies with no single Epistle slot this schema could
+   honestly fill; [holy_week_entries]'s own comment has the full argument
+   for why forcing a pair would be an uncited editorial choice, not a
+   Missal fact). Both entries: gated on colitur's own literal slug list
+   and [diffs] staying exactly [First_f; Gospel_f] (identity already
+   agrees on both). Every occurrence across the whole 2005-2050 fixture
+   falls in all 46 years without exception, on every one of these six
+   slugs -- Holy Week recurs every year by construction, so this is not a
+   coincidental count the way a collision-driven entry's would be.
+   Derived directly from the OCaml comparator's own failure output, not
+   hand-counted first and cross-checked after: C23 184 (46 x 4), C24 92
+   (46 x 2). *)
+let holy_week_answered_slugs =
+  [ "ef-passiontide-2-monday"; "ef-passiontide-2-tuesday"; "ef-passiontide-2-thursday";
+    "ef-passiontide-2-saturday" ]
+
+let holy_week_unanswered_slugs = [ "ef-passiontide-2-wednesday"; "ef-passiontide-2-friday" ]
+
+(* C25 (task 8, fix round 1, coordinator review, Important 3(b)) --
+   [verdict open], the SAME RG 69 gap C6 above re-opened, one hop removed.
+   RG 69 (docs/research/scan1.txt:625-631, word for word: "De dominica
+   infra octavam Nativitatis Domini... semper fit Officium cum
+   commemoratione festi forte occurrentis... nisi dominica incidat in
+   festum I classis") is unconditional: a Sunday landing 26-31 December
+   should keep ITS OWN Office (with the day's feast merely commemorated),
+   not the fixed weekday placeholder -- but colitur's [Temporal_ef] gives
+   both the SAME slug, undifferentiated by weekday (C6's own note, and
+   data/ef/expected-divergences.sexp's own C6 entry, have the full
+   account). Confirmed, not merely inferred,
+   that THIS entry is the direct, one-hop-removed consequence of that same
+   gap: every single one of these rows' colitur-side citation is "Tit. 3,
+   4-7" (the fixed weekday formulary), because [Lectionary_ef.readings]'
+   step 3 resumes the SAME wrongly-undifferentiated Sunday C6 already
+   names, not a second, independent cause. A genuine, precisely-diagnosed
+   [Temporal_ef] defect -- out of this task's own scope (a behaviour
+   change to a shared, multi-round-reviewed kernel-adjacent function) --
+   recorded here per the coordinator's own instruction ("record it
+   prominently... for a follow-up task"), not fixed and not left silently
+   failing. Gated on the literal 4-slug family (the only members of
+   [ef-christmas-1-<weekday>] this mechanism can ever reach -- Monday and
+   Tuesday of that same week are always claimed first by the Circumcision
+   or Holy Name's own DIRECT entries, so they can never fall through to
+   this specific step-3 chain) and [diffs] staying exactly
+   [First_f; Gospel_f] (identity already agrees -- this is a pure content
+   consequence). Derived directly from the OCaml comparator's own failure
+   output, not hand-counted first and cross-checked after: 57
+   (19 + 19 + 13 + 6, Wednesday/Thursday/Friday/Saturday respectively). *)
+let rg69_one_hop_slugs =
+  [ "ef-christmas-1-wednesday"; "ef-christmas-1-thursday"; "ef-christmas-1-friday";
+    "ef-christmas-1-saturday" ]
 
 (* [layer_c_reason l c diffs] returns the [data/ef/expected-divergences.sexp]
    [id] this row-pair's remaining (post Layer A/B) diff set belongs to, or
@@ -789,8 +852,18 @@ let layer_c_reason (l : row) (c : row) diffs =
        the reviewer: dropping the RG 91 entry 17 elevation in temporal_ef
        (Class2 -> Class4 on octave days 5-7) left BOTH differential tests green
        with C6 still reporting 138. The slug guard below already pins identity;
-       [Rank] only removed the ability to notice a rank regression. *)
-    && diffs = [ Slug_f ]
+       [Rank] only removed the ability to notice a rank regression.
+
+       WIDENED, task 8 fix round 1 (coordinator review, Important 3(a)/(b)):
+       [First_f; Gospel_f] joined the accepted set. Important 3(a)'s own fix
+       (tools/bootstrap_lectionary.ml's [nativity_octave_entries]) gives
+       these dates a real, Missal-verified citation for the first time --
+       and it does NOT converge with lectio's own (differently incomplete)
+       Sunday-walkback answer, so ALL 138 of this entry's rows now carry a
+       citation diff too, not merely the slug one. See data/ef/expected-
+       divergences.sexp's own C6 entry for why this entry's [verdict] is
+       now [open], not [colitur] -- Important 3(b)'s own RG 69 finding. *)
+    && subset diffs [ Slug_f; First_f; Gospel_f ]
     && List.mem c.slug nativity_octave_day_slugs
   then Some "C6"
   else if
@@ -858,7 +931,10 @@ let layer_c_reason (l : row) (c : row) diffs =
     diffs = [ First_f; Gospel_f ] && String.equal l.first "-" && String.equal l.gospel "-"
     && List.mem c.slug lectio_no_reading_slugs
   then Some "C20"
-  else if christmastide_fixed_slug_family c && subset diffs [ Slug_f; First_f; Gospel_f ] then Some "C21"
+  else if List.mem c.slug rg69_one_hop_slugs && diffs = [ First_f; Gospel_f ] then Some "C25"
+  else if List.mem c.slug lent_ember_slugs && diffs = [ First_f; Gospel_f ] then Some "C22"
+  else if List.mem c.slug holy_week_answered_slugs && diffs = [ First_f; Gospel_f ] then Some "C23"
+  else if List.mem c.slug holy_week_unanswered_slugs && diffs = [ First_f; Gospel_f ] then Some "C24"
   else None
 
 (* ---------------------------------------------------------------------- *)