| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
guard claim
Item 2(a): the RG 26 omission gate (rank = Class4) is extensionally
identical, over every candidate this engine actually constructs, to the
refuted "privilege_of loser = Ordinary" predicate it replaced -- confirmed
by enumerating the whole candidate space over 1583-9999, substituting the
old predicate back leaves 257/257 green and byte-identical output. Adds a
synthetic row: a temporal-origin, Ordinary-privilege Class3 loser, which
the two gates disagree on (RG26: Commemorate; the refuted gate: Omit,
since it never looked at rank). Reverting the gate to the old predicate
turns exactly this row red:
Expected: "Commemorate(Ordinary)"
Received: "Omit"
Item 2(b): the existing "RG33 boundary" row's comment claimed it proved
the RG26 branch's own (not (is_vigil ...)) guard. It does not: the vigil
sourced there is Class2, already excluded by the branch's rank = Class4
test alone, with or without the vigil guard -- verified by deleting the
guard, which leaves 257/257 green. Corrected the comment to say what the
row actually shows (RG31's "if impeded, commemorated" vigil mandate, a
real and separate fact worth keeping) and added a genuinely discriminating
synthetic row: a temporal-origin Class4 loser whose slug is ALSO
vigil-shaped, a combination RG 91 does not describe but Precedence.resolve
can still be asked to construct. Deleting the guard turns this row red:
Expected: "Commemorate(Ordinary)"
Received: "Omit"
Both reversions were applied locally to precedence_ef.ml, confirmed, and
reverted; the committed .ml is unchanged. Test count 257 -> 259, both new
rows green under the current, correct code.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Validation layer 5: fourteen hand-verified pins in test/test_golden.ml,
wired into the suite via test/test_colitur.ml. Each pinned day was computed
and checked against its RG citation before being written down, never
transcribed from `colitur day` output:
- Easter extremes 1598 (earliest, 22 Mar), 1666 (latest, 25 Apr) and 2038
(late-modern instance of the same latest date) -- all three independently
computed by hand via the Gauss/Meeus Gregorian Easter algorithm, not read
off Computus.gregorian_easter. extreme_years in test_validate.ml already
established 1598/1666 as the true 1583-2500 extremes (correcting a stale
1818/2038 comment); this file pins the resolved DAY there, not just the
date.
- Annunciation transfer, 25 March inside Holy Week (2016: 25 March is
literally Good Friday) and the double transfer with St Joseph (2008:
RG96's Attamen(a) claims Easter+8 for the Annunciation first, Joseph's
own RG96 walk continues past it to the next day).
- The RG96 Attamen(a) exception's own CONDITION pinned on both sides: 2057,
2007, 2012 (general walk suffices, lands before Easter, no exception) vs
2016/2024 (walk would cross Easter, Easter+8 fires).
- 2011-07-04, the Precious Blood transfer (Sacred Heart outranks it
outright on 1 July; the RG96 walk skips Visitation and a Sunday before
landing on 4 July).
- All Souls falling on a Sunday (2025, RG96 Attamen(b)) and Christmas
falling on a Sunday (2022, RG91 entry 1 -- no contest, since 25 December
is never an ordinary Sunday candidate in the EF temporal cycle).
- Holy Thursday's white amid violet Passiontide (2026, RG128(b)/RG122) --
the case colitur and lectio previously agreed was violet, so the
differential could never have caught it; only a golden pin or the
missalemeum oracle can.
- Advent/Lent Ember ferias commemorated when impeded (1900, 1902, RG24 +
RG109(e)) contrasted with IV-class ferias never commemorated (2026,
RG26) and RG111(b)'s Sunday rank floor (2009, 2026) -- all three assert
the actual displaced/excluded candidate is present in `omitted`, not just
that `commemorations` is empty, so a day with no losing candidate at all
could not pass vacuously.
2038 is deliberately NOT pinned day-by-day: register item F4 confirms
2038-03-06/08/09 are wrong (a lectio bootstrap-generator defect, not
fixable here). Only the Easter-week days, untouched by that bug, are
pinned; the exclusion is stated in the test's own comment, not silent.
Every weekday asserted was independently cross-checked against `date -d
<iso> +%A` (glibc, wholly outside this codebase) before being written down.
Precedence outcomes were traced against precedence_ef.ml's own band/
disposition/admit/transfer_target, not merely observed to look plausible;
where the primary text alone doesn't fully settle an outcome (the 2008
Annunciation/Joseph tie-break, both landing at RG91 table entry 11), the
test's own comment says so rather than overclaiming a citation.
Perturbation-tested: temporarily broke Holy Thursday's white-colour special
case in temporal_ef.ml, confirmed the golden test failed with a clear
day-and-field diff, reverted.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
M11 (the Nativity Octave's 26/27/28 December commemoration) previously
read "No RG paragraph found to support this" with verdict colitur.
Wrong on both counts, found by fix-round-1 review.
RG 67 ("dies infra octavam sunt II classis") and RG 109(c) ("de
diebus infra octavam Nativitatis Domini") both name this
commemoration; the 1962 calendarium's own text confirms it directly,
verbatim, under all three dates: "S. STEPHANI PROTOMARTYRIS, II
classis. / Com. octavae Nativitatis." (same pattern for John the
Evangelist and the Holy Innocents).
It is also a genuine regression the previous commit's own
commemoration-eligibility fix introduced: before that fix, colitur
happened to commemorate the day's generic Christmastide loser
unconditionally, which coincidentally kept presence agreeing with the
oracle even though the reasoning was wrong; the fix correctly stopped
that, but 26-28 December's own temporal candidate falls through to
[ferial_rank]'s Christmastide catch-all, which gives it Class4 -- the
wrong ferial rank for a day RG 67 says is II class. A naive fix would
also get RG 69 backwards: when this window falls on a Sunday, RG 69
names the SUNDAY as observed with a commemoration of THE FEAST ITSELF
(Stephen/John/the Innocents), never the generic octave placeholder --
so this needs a commemoration that attaches to whichever named feast
is observed, not one more competing RG 92-95 candidate the way
everything else in this engine is modelled. Out of this fix round's
own scope.
Verdict changed from colitur to open: neither side is confidently
right without that fix, and the previous "colitur is correct, nothing
to implement" framing was itself the risk under review.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RG109
Review round 1 (F1, Critical): the commemoration-eligibility fix from
the previous commit silently dropped II-class Advent and Lent Ember
ferias when impeded, while ordinary (lower-solemnity) ferias of the
same seasons kept their commemoration -- backwards on any reading.
Reproduced: 1900-12-21 (an Advent Ember Friday) lost its commemoration
entirely; 2026-12-21 (an ordinary Advent feria the same week) kept
its.
F2 (Important, the direct cause of F1): the branch's own justification
mis-stated RG 109 as an exhaustive list of the only temporal
circumstances that generate a commemoration. RG 109 is headed
"Commemorationes privilegiatae sunt commemorationes" and closes
"Omnes aliae commemorationes sunt commemorationes ordinariae" -- it
sorts commemorations that already exist into two HONOUR classes
(RG 108's differing liturgical hours), and says nothing about which
offices have the right to be commemorated at all. That right belongs
to Caput IV, "De feriis" (RG 21-27), never opened by the original
pass:
RG 24 (II-class ferias): "si vero impediuntur, commemorari debent"
-- if impeded, MUST be commemorated. Not optional.
RG 25 (III-class ferias): same mandate.
RG 26: every feria not named in 23-25 is IV class, and IV-class
ferias are NEVER commemorated -- the one exclusion RG 21-27 actually
states.
The omission branch is now gated on `rank = Class4` (RG 26) directly,
not on `privilege_of = Ordinary` -- the old gate happened to reach the
right answer for IV-class ferias (RG 26 also excludes them, for a
reason the old citation did not give) and the wrong one for II/III-
class Ember ferias RG 109(e) did not name by letter.
RG 109(e) itself is also corrected: its bare "feriis Adventus,
Quadragesimae et Passionis" previously excluded the Advent and Lent
Ember sub-days by analogy with (d)'s separate September carve-out.
RG 91's own TABLE needs an explicit "exceptis feriis Quatuor
Temporum" at entries 22 and 25 to keep Ember days from being double-
listed against their own entry 18 -- an exception that would be
unnecessary drafting if "feriae Adventus"/"feriae Quadragesimae" did
not already include their Ember sub-days by default. RG 109(e) carries
no such exception, so it is read at that same default, inclusive
scope: the Advent and Lent Ember ferias are privileged under (e), not
merely ordinary-but-commemorable. (d)'s own existence is unaffected --
September Ember days sit outside Advent/Lent/Passiontide under any
reading, so (d) remains necessary regardless.
Verified: 1900-12-21 and 1902-02-22 now correctly commemorate their
Ember ferias; 2026-12-21 and 1902-02-24 (ordinary ferias) unchanged.
Blast radius re-measured against the prior commit, 1900-2100: 102
civil days changed, every one an Advent/Lent Ember (or Ember-vs-
Joseph-collision) day regaining its commemoration, nothing else.
Validate re-swept exhaustively, 1583-9998: 0 failures.
F6: two privilege-boundary test rows that had been bent to expect the
bug's own output (Commemorate -> Omit) are restored to what RG 24/25
actually require, now Commemorate(Privileged) given the RG 109(e)
correction above -- these are the rows that should have caught F1.
F7: admit_cases had no witness for RG 111(b)'s rank floor that wasn't
already Class2, so reverting that filter only reddened the oracle
suite, never this file. Added a row (II-class Sunday, sole candidate
an ordinary Class3) that fails without the filter and passes with it.
F9: RG 122 cited alongside RG 128(b) for Holy Thursday's white --
states the same fact affirmatively ("Demum adhibetur color albus,
feria V Hebdomadae sanctae...") rather than as an exception to violet.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Validation layer 4: an oracle harness against missalemeum (Divinum
Officium data), independent of the lectio bootstrap chain colitur's
own sanctoral data comes from -- the only layer that can catch an
error inherited from that bootstrap, and the only one that can
validate commemorations at all (the lectio differential explicitly
excludes them, per its own header comment).
tools/extract_missalemeum_oracle.py shapes the fixture from lectio's
sources/snapshot.tar.gz outside the test (no JSON library in this
project's frozen deps, same reasoning test_differential.ml's own
fixture already documents). test/fixtures/missalemeum-ef-2026-2027
.txt (730 days, SHA-256 pinned and asserted) + its own .provenance
note record exactly how to regenerate it.
test_oracle.ml compares three axes the oracle actually supports:
rank, colour (SET MEMBERSHIP -- 14 of 730 days carry two colours,
e.g. rose+violet on Gaudete/Laetare, which independently vindicates
this project's own rose reading against lectio's violet-only one,
recorded in the register), and commemoration presence/count. Slug
identity is deliberately out of scope (needs a title->slug mapping,
the data audit's own business, not the automated comparator's).
Of 730 days, 688 matched cleanly outright. The remaining 42 are all
named in data/ef/expected-divergences-missalemeum.sexp (14 cited
entries, M1-M14): most are genuine primary-source-confirmed findings
this task adjudicated and fixed in the two preceding commits (RG 33,
RG 109/111, Holy Thursday's colour); the rest are real, cited,
deferred feature/data gaps (RG 91 entry 27's BVM-Saturday office, RG
110's inseparable Peter/Paul commemoration, four sanctoral entries
missing from lectio's own source) or genuine oracle-side artifacts
-- honestly verdicted against whichever side this task's own
primary-source research actually backs, never defaulted to colitur.
One entry (M13, St Joseph vs the Friday of Passion Week 2027) is
verdict open: adjudicated as unresolved after real search effort,
not guessed past.
The data audit: every sanctoral entry the comparison flagged was
hand-checked against the 1962 calendarium, plus a 20-entry
deterministic random control sample (seed 20260812) drawn
independently of the flagged set. The control sample caught two
entries (benedict, frances-rome) marked Commemoration_only in the
bootstrapped data when the primary calendarium lists them as plain
III-class feasts with their own Office -- traced to lectio's own
source, not fixable here, and reported as a signal (10% of a random
sample) rather than a blanket claim. Coverage recorded honestly in
docs/research/rules-register.md's own three buckets: confirmed by
oracle (200/322), confirmed by hand (23/322, 2 of them wrong),
unverified (115/322) -- the unverified bucket stated explicitly
rather than left implicit.
Harness teeth demonstrated and reverted (not committed): a fixture
rank/colour edit on a previously-clean day fails both the checksum
pin and the no-unexplained-differences assertion independently; an
expected_rows drift on the allow-list fails the citation-count
assertion. Both captured with their exact failure messages, both
reverted before this commit.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RG 128(b) excepts "Missa sive Chrismatis sive in Cena Domini feria V
Hebdomadae sanctae" from the Septuagesima-to-Easter-Vigil violet span
as a whole-Mass exception (unlike Palm Sunday's blessing/procession,
which the same sentence carves out as only part of that day) -- so
Holy Thursday's Mass is white, not violet. temporal_ef.ml fell through
to the generic Passiontide ferial path (season_colour Passiontide =
violet) for this date, since [named] has no entry for the Triduum at
all (rank was already correct via [privileged_feria]; only colour was
wrong).
Found via Task 16's missalemeum oracle comparison: every other Triduum
day's oracle colour set includes violet as an option (Good Friday
"bv", Holy Saturday "vw" -- already-flagged, deliberately deferred
per-action gaps, register §3b), but Holy Thursday's is white alone.
This is a genuine NEW divergence from lectio too (lectio has no such
exception and still prints violet for Holy Thursday) -- added as
Layer C11 to the differential harness and data/ef/expected-divergences
.sexp, 46 rows (2005-2050), rather than silently changing what that
harness accepts.
Also corrected a stale comment on season_colour claiming white's RG
paragraph was never pinned -- RG 119 was found and cited in the
register on 2026-08-11 but the correction was never copied into this
file's own comment.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Task 16's missalemeum oracle comparison (2026-2027) surfaced three
distinct precedence-engine bugs, each confirmed against the primary
1962 Missale Romanum text and cross-checked against the oracle:
1. RG 33 (vigil omission) was transcribed backwards. The primary text
reads "Vigilia II AUT III classis penitus omittitur" (a II OR III
class vigil is entirely omitted on any Sunday or I-class feast) --
not "I aut II" as the register and this module's own rank check
(Class1 || Class2) previously had it. I-class vigils (Nativity,
Pentecost) can never actually lose in this engine (band always
ranks them above anything that could coincide with their dates), so
the Class1 half was dead code; the real bug was that Class3 (St
Lawrence's vigil, the sole III-class vigil) was MISSING, so it fell
through to an ordinary commemoration instead of RG 33's mandatory
omission whenever 9 August landed on a Sunday.
2. RG 93/95/109/113 read together: an ordinary, non-privileged
TEMPORAL-cycle office (a bog-standard green-season feria, a plain
Advent/Lent Ember day, a Minor Rogation day) has no standing to be
commemorated at all when impeded -- RG 109's six-item list is a
CLOSED set of the only temporal circumstances that generate a
commemoration, not a floor under which anything ordinary still
gets one. The engine previously commemorated the losing feria
itself in this situation; confirmed wrong against ~190 independent
missalemeum days (2026-2027) showing zero commemorations for the
exact shape. Vigils are explicitly excluded from this new rule --
RG 31/32 give them their own "if impeded, commemorated" mandate,
independent of RG 109's list.
3. RG 111(b)'s Sunday admission slot is restricted to "de festo II
classis" -- a rank restriction, not merely the best available
ordinary candidate. A III/IV-class ordinary saint competing for a
II-class Sunday's single commemoration slot had no such restriction
applied before; confirmed wrong via St Hyginus (11 Jan, Class3)
losing to Holy Family, which missalemeum shows entirely displaced,
never commemorated.
All three fixes are pure disposition/admission changes -- they never
touch band, so the observed day (season/slug/rank/colour) is
unaffected in every case; the lectio differential (which never
compares commemorations) stays green untouched. test_precedence_ef.ml
updated throughout: two previously-wrong test expectations corrected
(the RG33 III-class-vigil boundary row, the IV-class-feria totality
row), three privilege boundary rows sharpened from "commemorated but
unprivileged" to "omitted entirely" now that temporal+ordinary
means omission, and a new contrast row added to keep the SANCTORAL
side of rule 2 covered separately from the TEMPORAL side.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix round 1 on Task 15's review, four findings closed.
C1 and C6 previously gated on calendar date alone -- unlike every other
Layer C entry, which also checks the actual colitur slug that won. A
future sanctoral change producing an unrelated celebration with the same
diff shape (e.g. a promoted 29-31 December saint landing on C6's own
[Slug_f; Rank]) would have been silently absorbed under a citation that
has nothing to do with the real cause. Both entries now require, whenever
Slug_f is in the diff set, that colitur's own slug be a member of an
explicit closed list (jan_6_13_slug, nativity_octave_day_slugs) -- the
same style advent_feria_slug already used for C4. Constructed the exact
substitution via a temporary (unshipped) overlay Replace directive
promoting thomas-becket to Class1 Feast White; confirmed the tightened
predicate refuses to absorb it (46 unexplained rows, one per year); reverted.
C10 gained the diff-field-subset restriction every other entry already
had (it previously fired on the two pinned dates regardless of which
fields differed).
Fixed a stale doc comment ("SIX" columns; field has five constructors).
The fixture's provenance note now records its SHA-256, and the suite
asserts it as its own first test case (test_fixture_checksum), so a
hand-edited or partially re-copied fixture fails loudly instead of
silently becoming an unlabelled snapshot. Computed via the system
sha256sum binary through Sys.command plus a redirected-output temp file
-- no new library dependency, not even unix. Constructed a one-byte
fixture edit and confirmed the checksum test fails independently of the
row-comparison test; reverted.
237/237 tests green, clean-build verified.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Validation layer 3 (design spec's five): compares colitur's real day-by-day
EF output against lectio (sibling project, Go), 2005-2050, one line per
civil day. Of 16801 day-pairs, 11206 already agree on the seven leading
columns; the 5595 that don't resolve into exactly 25 distinct field-diff
signatures, all triaged.
Three strictly separate layers, per the controller's ruling (the brief's
single flat allow-list assumed a handful of differences, not 5595):
- Layer A (test_differential.ml, norm_season/norm_slug): vocabulary. An
explicit, closed table of naming synonyms with no liturgical substance
(lectio's easter/christmas vs colitur's paschaltide/christmastide; a
handful of slugs that are two names for the identical office). No
wildcards -- every entry is a literal string pair.
- Layer B (strip_epiphany_index): numbering. The one slug family whose
lectio/colitur index offset is not a constant (Time-after-Epiphany week
numbering, register 3c#5) has its embedded digit stripped to a common
form on both sides before comparing; rank and colour stay fully compared.
- Layer C (data/ef/expected-divergences.sexp): the cited allow-list. Ten
genuine liturgical disagreements, each citing its RG paragraph and
naming which engine is right (always colitur, verified against the
Missal/register, never against lectio's own behaviour). This is the
only layer permitted to cover a difference in rank, colour, or which
celebration is observed. Five extend or restate register 3c's already-
documented divergences (season boundary, Sunday I-class, Advent Ember
ferias, Rogations); five are new, found and adjudicated in this task
(Lent Ember days, the Nativity Octave, Ember-day-vs-saint precedence,
the St Joseph transfer off a Lent Sunday, and the 2011 Sacred Heart /
Precious Blood / Visitation collision). expected_rows on each entry is
an exact regression pin, asserted by the test, not documentation.
13 January (register 6's long-open "Baptism of the Lord" item) is
confirmed empirically fixed already -- Task 11's sanctoral wiring closed
it before this task started -- so it is not allow-listed; the only
residual difference there is the season boundary already covered by C1.
Two stated limits carried from the brief (commemorations are not
comparable; lectio's own EF oracle asserts season only, 2025-2026 only,
so a rank/colour difference is not presumptive evidence against colitur)
plus a third found during this task (the week column is a display
convention on both sides, not a liturgical fact, and is not compared at
all) are documented in the test file's own doc comment.
Fixture: test/fixtures/lectio-ef-2005-2050.txt, committed as plain text
(1.4 MB), generated by lectio commit 2386a45; provenance recorded in the
sibling .provenance file. Colitur's side is recomputed fresh from the
library on every run, through the same Calendar/Rite_ef pipeline
`colitur day` uses, not the compiled binary.
Proved the harness has teeth by two reverted perturbations: a genuine
colour difference injected into a fully-covered fixture row fails the
"no unexplained differences" check with the exact mismatched row printed;
a one-row drift in an allow-list entry's expected_rows fails the count
check independently, showing it is not merely a duplicate of the first
assertion. 236/236 tests green, clean-build verified, deterministic
across OCAMLRUNPARAM=R.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Three carried items from Plan 2's parked rulings, closed:
1. Slug uniqueness moves from a 200-sample QCheck property scoped to
one rite (test_temporal_ef.ml) into Validate's own "slugs" check, so
every consumer gets it. The resumed-Sunday exemption that property
carried is dropped, not weakened elsewhere: Plan 2 verified zero
duplicate slugs domain-wide (all 8 416 years), and by construction a
resumed Sunday only ever backfills a week number Septuagesima cut
short that same liturgical year, so it can never repeat a number
that year's own January Sundays already used. The now-redundant
property and its is_resumable_sunday_slug helper are removed from
test_temporal_ef.ml; test_validate.ml's own domain-wide property
covers the same ground for every consumer.
2. The anchors-erosion guard (Plan 2: deleting entries from a rite's
anchors list left the whole suite green) is implemented, but not in
Validate. Which of a rite's named days are Easter-derived is
knowledge only the rite's own `named` function has; Rite.t
deliberately exposes only `temporal` and `anchors`, never `named`,
so a rite-agnostic Validate has no ground truth to check anchors'
completeness against. Hardcoding an Easter offset, or even Easter
itself, would smuggle Western/Gregorian-specific knowledge into code
meant to also serve a future Julian-reckoning rite; rediscovering
"named-ness" structurally from `temporal` alone is unsound for EF,
since most ordinary Sunday/feria slugs from Septuagesima onward are
also constant-offset-from-Easter by construction. The guard is
therefore EF-specific and lives in test_temporal_ef.ml, discovering
the Easter-derived slug set mechanically (scanning a window around
Easter and keeping whatever `named` answers Some for) rather than
hand-copying either named's or anchors' own offset list, then
asserting completeness against the real anchors for the domain's
Easter extremes (1598, 1666) plus an ordinary year. A negative
fixture proves the guard has teeth, matching Plan 2's exact
regression (anchors missing "ef-ascension" reports it, and only it,
as missing).
3. test_validate.ml's extreme_years comment claimed 1818/2038; verified
against Computus.gregorian_easter directly, the domain's actual
Easter extremes (1583..2500) are 1598/1666. Corrected.
Verification: the full 1583..9999 domain sweep (233 tests via dune
test's 200-sample default, plus a manual full sweep) reports exactly
one failure -- the known, already-pinned year-9999 season-truncation
case -- and zero occurrences of the new "slugs" check anywhere in the
domain. Deleting "ef-ascension" from the real anchors list (reproducing
Plan 2's regression directly) is caught immediately by the new EF test
and, confirmed empirically, invisible to Validate's own full property
sweep -- direct evidence for why item 2 cannot live in Validate.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Task 11's re-review found that Precedence_ef.transfer_target's
Annunciation/Easter condition -- and RG 96's general walk more broadly --
is correct only because Temporal_ef happens to make every day from
Easter Sunday through Low Sunday (Easter+0..+7) blocking (Class1/2):
Easter and Low Sunday via named, every day between via
privileged_feria. Nothing enforces that; a future edit narrowing the
octave's blocking range could let a translated feast land inside it,
silently.
Add a test for the consequence, not the mechanism: over 2005-2050 (real
data, the project's own differential window), no day in [Easter,
Easter+7] is ever a transfer's target, checked against both
transferred_in and transferred_out's own recorded target dates. Plus a
direct, deliberately-constructed case that starts the RG 96 search
inside Holy Week itself (Holy Thursday 2026, real occupant), pinning
that the search lands on Easter+8 -- and a companion test that reads the
real occupant's own rank across the whole [origin+1, Easter+7] span to
confirm the search genuinely walked through the window rather than
stopping early for an unrelated reason.
Mutation-verified: narrowing privileged_feria's post-Easter range (6 to
4 days) makes all three new tests fail, including a real natural
occurrence the fixture did not anticipate (St Joseph, 19 March,
transferring into Easter+5 in 2008/2035/2046) -- not just the
deliberately constructed Holy Week case.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Widen Validate.run to take the rite's sanctoral layer alongside the rite
itself (Calendar.year needs both), and add five checks over the fully
resolved liturgical year, on top of the existing temporal-only pass:
- observed: a day's observed celebration never also appears among that
same day's own commemorations/omissions.
- lost: no sanctoral entry is silently dropped. Per slug, the number of
times it is actually sighted (observed + commemorations + omitted,
summed over the year) must never fall below the number of times its
own Date_spec resolves within the year's span -- also fires if
resolving the year raises at all, the most total form of loss.
- duplicated: the same per-slug count must never exceed the number of
Date_spec resolutions either. Deliberately NOT "no slug appears
twice": a fixed date can legitimately resolve twice in the ~20% of
liturgical years whose 371-day span reaches it on both ends (30
November/St Andrew is the worked example in validate.mli).
- unconverged: no day's omitted reason indicates Calendar's placement
pass hit its round guard before reaching a fixed point.
- admission: the rite's own rules.admit is a fixed point on what it
already admitted -- the rite-agnostic form of "the admission limit
was not exceeded" available without embedding a rite's own numeric
caps (RG 111's, for EF) into kernel code.
Each check has a dedicated negative fixture in the synthetic rite
(test_validate.ml), hand-traced against Calendar's actual resolution
mechanics before writing the assertion, and verified to fail for the
right reason against the code before this change. One pair
(unconverged/duplicated) is not fully independent: hitting the round
guard genuinely also trips duplicated, a real consequence of Calendar's
own accounting once a candidate is simultaneously sighted at its
permanent natural date and wherever the last placement round left it --
documented in guard_rules's own comment, not papered over.
test_validate.ml's ef_rite/run now use the real Rite_ef.context and the
real bootstrapped data/ef layer (Precedence_ef and the sanctoral
bootstrap did not exist when this scaffolding was first written) rather
than the earlier placeholder rules. Validate is clean across the whole
1583..9999 domain against real EF data except the one already-documented
year-9999 truncation case (test_year_9999_does_not_raise).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The register was transcribed faithfully but was itself wrong: RG 96's
Attamen (a), primary-source-verified from the scans and now corrected
in the register, reads 'festum Annuntiationis B. Mariae Virg., quando
est transferendum post Pascha, transfertur ... in feriam II post
dominicam in albis' -- the Monday-after-Low-Sunday seat applies ONLY
'quando est transferendum post Pascha', when the feast is to be
transferred PAST EASTER. The unconditional transcription made the
exception fire on every impeded Annunciation regardless of cause.
transfer_target now computes the general RG 96 target first, for every
candidate, and overrides to the Monday after Low Sunday only when that
general target itself falls after Easter Sunday -- testing the rubric's
own condition directly rather than re-deriving a date-proximity rule
from first principles.
Confirmed against three real years the review named: 2007, 2012 and
2057 all previously sent the Annunciation to Easter + 8 (16 April,
16 April, 30 April respectively) when the correct, now-produced target
is the next free day before Easter (26 March in each case -- Passion
Sunday in 2007/2012, Lent III Sunday in 2057). Verified with actual CLI
output for all three, before and after.
Also cites RG 96 Attamen (b), the same primary-source passage, as the
direct authority for All Souls' own move to the following Monday when
impeded by a Sunday -- previously inferred from RG 91 entry 8's
parenthetical plus the general walk, which happened to produce the
right date; now stated directly.
Rewrote the two existing Annunciation unit tests, whose synthetic
occupants no longer trigger the (now correctly conditional) exception,
and added a 2057 regression test using the real Temporal_ef.temporal as
occupant plus a real-data cram pin -- both mutation-verified against
the unconditional reading.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With data/ef/adjustments.sexp's one Suppress directive replaced by an
empty directive list, colitur day 2026's stdout was byte-identical for
all 365 days and the whole suite still passed: ef-nativity-vigil already
outranks vigil-of-christmas in the RG91 contest regardless of whether
the overlay ever ran, so the suppression's only observable effect is
keeping vigil-of-christmas out of 24 December's Liturgical_day.omitted
-- a field no cram assertion reads. test/dune's own dep on
adjustments.sexp pinned the file's existence, never its content.
Confirmed the gap two ways: reverting the overlay to (directives ())
makes this new test fail (24 December's omitted gains
vigil-of-christmas), and confirmed test/cli.t's whole cram suite still
passes under that same mutation -- cram is structurally blind to this,
not merely undertested.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
search_from could walk up to 400 days past origin before Calendar's own
~start ~stop clamp is ever consulted, and nothing stopped it probing
occupant on a date past 31 December 9999 -- occupant chains through the
real EF rite's temporal, which calls Computus.gregorian_easter, not
total outside 1583..9999 (it Date.makes and failwiths on Error).
Not reachable with the shipped sanctoral data alone, but reachable
through the project's own primary extension path: an overlay adding an
I-class feast on 25 December leaves nothing but Class2 Nativity-octave
days for the rest of civil year 9999, so the unguarded search reached 1
January of year 10000 and crashed there with 'computus: year 10000 out
of range 1583..9999'. 9999 is an in-range year and the kernel's contract
is 'never raises on in-range input'.
search_from now also stops, without probing occupant again, once it
passes Date's own domain ceiling -- the same 'return a finite date, let
Calendar's own out-of-range handling record it, never pretend to have
found something admissible' contract the existing step-count guard
already follows.
Two new tests, both mutation-verified to actually reproduce the crash
when the guard is removed (see the task report): a precedence_ef.ml unit
test using the real Temporal_ef.temporal as occupant (a synthetic
occupant can never discriminate this, since it never calls Computus
itself), and a Calendar-level integration test reproducing the exact
overlay-based scenario the review found.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Assembles Rite_ef.context (lib/rites/rite_ef/rite_ef.ml[i]): temporal,
anchors and vocab from Temporal_ef; rules from Precedence_ef's band,
disposition and admit; season_runs = Vocab_ef.seasons; transfer_target
newly implemented here.
transfer_target (RG 96): the next following day that is not I or II
class, with the Annunciation's own exception (Monday after Low Sunday).
Terminates by a structural step bound on its internal search, independent
of Calendar's own round guard, which bounds rounds across a year, not one
call's walk; documented as an obligation on rite.mli's transfer_target
field, which did not previously state it.
Fixes the vigil-naming mismatch Task 7's review predicted: the sanctoral
bootstrap names its vigils with a vigil-of-X prefix (lectio's own
convention), while Precedence_ef's is_vigil only recognised the temporal
cycle's own X-vigil suffix. Both are now recognised, fixing RG 91 entries
21/26 and RG 33's vigil omission for the four affected celebrations.
Verified by unit test and by mutation-testing the fix (reverting it fails
exactly the new rows) and against real output across several years.
Suppresses data/ef/sanctoral.sexp's vigil-of-christmas via a new overlay,
data/ef/adjustments.sexp: it is the same celebration as the temporal
cycle's own ef-nativity-vigil, both dated 24 December.
colitur day <year>: one line per civil-year day, temporal and sanctoral
fully resolved through Layer, Overlay, Precedence_ef and Calendar -- the
first CLI path exercising the whole Plan 3 pipeline against real data.
Verified the All Souls transfer chain (Tasks 7-8-11) end to end against
real output for both a Sunday year (2025, lands on 3 Nov) and a
non-Sunday year (2026, observed directly on 2 Nov).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Convert lectio's tridentine-calendar.ini (322 entries) into
data/ef/sanctoral.sexp via a validating OCaml converter, tools/
bootstrap_sanctoral.ml, rather than a hand-written script: every field is
built through Slug.of_string, Colour.of_string and Vocab_ef.rank_of_string,
so the emitted sexp is valid by construction.
Two conversion decisions, both documented rather than buried:
- subject defaults to Subject.Saint, overriding Celebration.make's kernel
default of Subject.Temporal, for the 316 entries with no explicit class;
- rank = commemoration maps to status = Commemoration_only with an inferred
Class3 (not a citation -- it is what the 1960 reform reduced most simple
feasts from), recorded as an open item in the rules register for the
oracle to adjudicate.
Every celebration is tagged layer = Precedence_ef.universal_layer, the
provenance id RG 91's band classifier reads to tell the universal calendar
from proper/indult data.
The generated file carries a provenance header: source path, its SHA-256,
and the UTC conversion date, so re-bootstrapping against a newer lectio is
reproducible and diffable. Output is byte-identical across runs.
test/test_sanctoral_ef.ml loads the file through Layer.load and checks the
counts independently derived from the source INI (322 entries, 114
Commemoration_only, 12 Class1, no Subject.Temporal, every date resolves in
a leap year), plus two named spot-checks against the INI's own text -- one
entry with an explicit class field, one commemoration -- so a passing count
cannot hide the wrong 322 entries having been converted.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds test_ii_class_sunday_privileged_witness_admitted_end_to_end,
proving RG 111's "II-class Sundays: one, dropped if a privileged one is
due" clause through the real pipeline (band, disposition, privilege_of,
admit together), not only at the isolated admit level.
RG 109(b) is the one privileged category that is sanctoral and
rank-based rather than temporal-origin: a Commemoration_only, Class1
sanctoral candidate (the same shape disposition_cases already uses as
(b)'s witness) is held out of the band contest entirely by
Precedence.resolve, so it always loses regardless of what wins the
day. Paired with an ordinary sanctoral saint, both losing to an
ordinary Class2 Sunday, gives observed_is_sunday = true with one
privileged and one ordinary loser due at once, with no new fixture.
Documents, and mutation-confirms, what this fixture does and does not
prove: RG 109(b)'s only reachable witness is Class1 (the highest
dignity), so the privileged loser here is unavoidably also the
higher-dignity one, and the admission outcome alone cannot distinguish
"privilege overrides dignity" from "dignity alone picked the same
winner" -- that distinction remains admit_cases's job. What this test
does prove, and what a category-(b) regression in privilege_of breaks,
is the privilege tag itself surviving the real pipeline end-to-end.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
disposition's Class1 branch was unconditional on rank, so an impeded
I-class Sunday (Advent/Lent/Passiontide/Low Sunday) currently
transferred like a feast. RG 95 (register lines 323, 363) restricts the
right of translation to I-class FEASTS -- RG 91's own table lists
Sundays as a separate row (entry 6, line 332) from feasts (entries
11-13, lines 337-339) -- and RG 109(a) (line 374) lists "of a Sunday"
as a privileged commemoration category, which presupposes an impeded
Sunday stays put rather than moving to another day.
Excludes is_sunday_slug losers from the Transfer branch so they fall
through to the existing Commemorate (privilege_of loser) branch, which
already tags them Privileged via RG 109(a) with no further change.
Fires on real dates in the 2005-2050 differential window: 24 December
on Advent IV in 2023, 2028, 2034, 2045; 8 December on an Advent Sunday
in 2024, 2030, 2041.
Verified no previously-passing Transfer assertion used a Sunday-shaped
loser (grepped test_precedence_ef.ml, test_calendar.ml -- which uses
its own synthetic rite unrelated to Precedence_ef.disposition --
test_colitur.ml, test_validate.ml); confirmed by mutation-reverting the
fix and observing exactly the one new test fail, nothing else.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Precedence_ef.privilege_of classifies a commemoration candidate against
RG 109's closed list of privileged commemorations (a Sunday; a I-class
day; a day within the Octave of the Nativity; a September Ember day; a
feria of Advent, Lent or Passiontide; the Major Rogations in Mass), read
entirely off the candidate's own rank/slug/origin, no context needed.
Major Rogations (f) is left unimplemented rather than guessed: no
producer for the Major Litanies exists anywhere in this codebase yet.
disposition's two Commemorate sites now call privilege_of instead of
Task 8's interim_privilege placeholder, which is removed entirely
(binding and .mli export both gone).
Precedence_ef.admit applies RG 111's four admission counts, keyed on
the observed day's own class and Sunday-ness: a I-class day admits none
except one privileged commemoration; a II-class Sunday admits one, but
a privileged commemoration due displaces any ordinary one regardless of
dignity; any other II-class day admits one by dignity alone, with no
such override; III/IV-class days admit at most two by dignity. Ties
break on slug, matching Precedence.compare_by, so the admitted set
never depends on input order. Every admitted candidate is a value taken
unchanged from the input list, never rebuilt, so Precedence.resolve's
physical-equality-based dropped/omitted accounting keeps working --
closing a note left open since Task 2.
Tests: RG 109 category rows (one per reachable category, plus boundary
rows proving Advent/Lent Ember days and Minor Rogations are correctly
excluded), RG 111 admission rows checked on slug identity rather than
count, an order-independence check, and three end-to-end integration
tests proving the admission limit's drop lands in resolution.omitted
rather than vanishing.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Precedence_ef.disposition decides the loser's fate in an occurrence: a
Commemoration_only celebration is always commemorated (it can never win or
transfer); a I- or II-class vigil impeded by any Sunday or a I-class feast
is entirely omitted (RG 33), checked before the generic rule below or the
Nativity/Pentecost Vigil could wrongly transfer; any other I-class loser
transfers (RG 95 -- only I class has the right of translation); everything
else is commemorated, with the admit-or-omit decision left to RG 108-111's
admission count (Task 9). RG 94 needs no code: resolve always compares a
loser against the day's actual winner, never against a departed sibling, so
no commemoration can ride along with a transferred feast in this design.
This is the branch that completes Task 7's carried All Souls fix: once it
loses to an occurring Sunday, its untouched Class1 rank routes it to
Transfer via the generic rule, not a special case. Landing on 3 November is
Rite.transfer_target's job, not wired up yet.
Commemorate carries an interim Precedence.Ordinary privilege pending Task
9's RG 109 implementation, exposed as interim_privilege for that task to
replace.
Table-driven tests cover each rule, including RG 33's boundary from both
sides and a Commemoration_only loser that is also Class1 and vigil-shaped
to pin the branch ordering. Mutation-tested: disabling RG 33, either
direction of RG 95's rank condition, or the Commemoration_only priority
check each fail exactly the rows built to catch them.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Review of 436ba75 found two calendar defects and one coupling risk.
Entry 8 (All Souls) dropped register line 334's own qualifier,
"(yields to an occurring Sunday)" -- it returned 8 unconditionally,
so on 2 November falling on a Sunday (2025, 2031, 2036, 2042 in the
2005-2050 differential window), All Souls incorrectly outranked and
observed over the Sunday. 2 November is always Time_after_pentecost
and never coincides with any other entry's own date, so the only
rival this exception ever has is an ordinary entry-15 II-class
Sunday; on such a Sunday, band now returns one worse than entry 15's
own value rather than the literal 8 -- strictly worse, not merely
different, since an exact tie would fall to Precedence.resolve's slug
tie-break, which for "ef-all-souls" against a Sunday slug would make
All Souls win the tie anyway. Entry 8's own rank is untouched, so
Task 8's disposition still sees a true I-class candidate to transfer.
Entry 14 (Feasts of the Lord, II class) added a universal-layer
restriction the register's line 341 does not carry -- contrast entry
16 at line 342, which explicitly says "not of the Lord"; RG 37c
(register line 393) also speaks of "II-class feasts of the Lord"
with no universal qualifier. Dropped the layer test: a proper or
indult feast of the Lord now bands 14, not 19/20.
Also exposes vigil_suffix and ember_prefixes from precedence_ef.mli,
matching universal_layer/indult_prefix's existing "colitur convention,
not an RG citation" treatment -- both were previously private literals
duplicated in the test file, so a rename of temporal_ef.ml's slug
format could have drifted silently past both sides agreeing with each
other. Two test rows now build their candidate from Temporal_ef.temporal's
own real output (entry 18's Lent Ember day, entry 21's Ascension Vigil)
instead of a hand-typed slug, closing that specific coupling.
Adds the three cheap rows review named as closing the remaining
unpinned guards (a temporal-origin Class1 candidate on an unnamed
date; a universal Class2 vigil of the Lord; a non-universal Class3
vigil), plus a resolve()-level test proving the Sunday is actually
observed over All Souls, not just that band returns the right integer
in isolation.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Precedence_ef.band transcribes RG 91's 28-entry Table of Precedence
(rules-register.md §4) for the EF rite: given a day's context and a
candidate celebration, returns the table's own entry number, 1-28
(I class 1-13, II class 14-21, III class 22-26, IV class 27-28); lower
wins. Every branch carries its entry number and register citation in a
comment, checked in the table's own numeric order.
Two entries are transcribed as the register states them even though
they invert the pattern the rest of the table follows: at III class,
23 (particular calendars) outranks 24 (universal), the reverse of how
11/12 and 14/16/19/20 rank a universal feast ahead of a proper one at
I and II class.
Sanctoral-origin, layer-decided entries (11-13, 14/16/19/20, 23/24)
follow the brief's structural insight: a celebration whose layer is
not the universal base is an overlay -- proper, or indult if its
layer id also carries the indult prefix. Neither the universal-layer
id nor the indult prefix is an RG citation; both are colitur's own
data-modelling convention, exposed from the module so whichever task
loads the real EF sanctoral overlays can align to them.
Vigils (21, 26) are read off the temporal cycle's own -vigil slug
suffix rather than gated on origin, since a II/III-class vigil can be
either temporal-origin (Ascension, already produced by temporal_ef) or
sanctoral-origin (a saint's vigil, no task has loaded yet); Ember days
(part of entry 18) are read off temporal_ef's own ember slug prefixes
rather than re-derived, since the September anchor is independently
flagged there as one of the more contested dates in the calendar.
A candidate shape the table has no row for (e.g. a Class1 vigil that
is not Nativity or Pentecost, or a Class4 candidate marked as a vigil
-- RG 91 has no IV-class vigil either) returns a dedicated unclassified
sentinel (max_int) rather than being folded into a same-rank entry it
does not belong to.
test_precedence_ef.ml is table-driven: one Alcotest.test_case per RG
91 entry (55 rows total, several entries covered by more than one
named day so a single missed offset cannot hide behind a passing
sibling), each date computed from Computus.gregorian_easter rather
than hand-typed, so an arithmetic slip cannot pass by accident.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Four findings from Task 6 review, addressed on top of f15e44d.
1. transferred_out was a single Date.t option, so when RG 97-98 collides
three or more feasts on one date (more than one loser), only the last
one Hashtbl.iter happened to visit survived -- a genuinely lost move,
and which one survived depended on OCaml's hash seed
(OCAMLRUNPARAM=R), an environment read a kernel invariant forbids.
RG 97-98 says coinciding feasts transfer "in order" -- plural -- so
the type was wrong, not the fixture: transferred_out is now
(Celebration.t * Date.t) list. transferred_in stays a single option,
deliberately: a day receives at most one arrival (RG 96 sends each
departure to the next non-I/II-class day, and the first to arrive
occupies it). The per-day list is canonicalised (sorted by target
date, then slug) after accumulation, the same fix layer.ml already
applies to its own date-bucket index and for the same reason.
Verified clean across 15 runs under OCAMLRUNPARAM=R; disabling the
canonicalisation step showed the raw order genuinely flip between
seeds, confirming the fix is load-bearing.
2. Every deferred candidate in the fixture was the same rank, so
compare_deferred's band branch was unreachable and reversing it broke
nothing -- the RG 97-98 test was pinning slug order, not band order.
The fixture now has three ranks (Hi1 outranks Hi2, both transfer,
both outrank Lo), with slugs chosen so band order and slug order
disagree. Reversing the band comparison now fails the test on
"higher-band loser claims 2 Feb first", received the wrong slug
instead.
3. A transfer_target free to name any date could place a candidate
outside the liturgical year's own start/stop bounds: invisible to
year/build_day, so it would be observed nowhere and, since its
origin's re-resolution would report it as settled, omitted nowhere
either -- genuinely gone, contradicting calendar.mli's "never
silently dropped". place_transfers now checks the range on every
placement and routes an out-of-range one to a permanent-exclusion
table instead of assignment, with its own cited omitted reason.
4. Precedence.resolve folds Transfer and Repose into one deferred case,
and place_transfers routed all of it through transfer_target (RG
96's search), which is only correct for Transfer. Repose is RG
100-102's repositio, a distinct rubric this module does not
implement. Documented rather than split into a second mechanism:
nothing in the EF ruleset returns Repose (design spec section 1.3,
"declared, not exercised"), so the gap is latent, not a live bug.
Two new tests (origin records every departure; transfer target outside
year is recorded not lost); the RG 97-98 test's fixture and assertions
rewritten for finding 2.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Calendar.year now runs a placement pass after resolving every day: each
deferred candidate (RG 95's I-class-only right of translation, via
Precedence's Transfer disposition) is placed on the next day the rite's
new Rite.t.transfer_target names as admissible, transferred_in/out are
set on the two ends of the move, and the whole year is re-resolved to a
fixed point, bounded by a hard max_transfer_rounds = 64 guard.
transfer_target is rite-supplied rather than a generic search Calendar
drives itself: RG 96's 'not I or II class' is not derivable from band or
disposition alone (RG 91's own table lets a universal I-class feast
outrank an ordinary Sunday in a raw contest, yet RG 96 forbids landing a
translation there regardless), and the search's starting point is
rite-specific too (the Annunciation exception). It takes an occupant
callback exposing what Calendar currently resolves as observed on any
date, so the rite never has to re-implement occurrence resolution.
Two correctness properties drove most of the design:
- A candidate's permanent natural loss at its own origin (the layer entry
never moves) is rediscovered every round; left unfiltered this
oscillates a placed candidate between two dates forever, since its own
rank makes it look 'occupied' to a fresh search from its origin. Both
the round loop's gather and the final per-day omitted accounting filter
this out, keeping only sightings that are either brand new or losing at
a candidate's *current* target (a fresh RG 97-98 bump).
- RG 97-98's sort has to actually decide something, not just happen to
agree with Precedence.resolve's own tie-break next round: a
claimed-this-round overlay lets earlier-processed candidates in one
round block later ones in the same pass, so two coinciding I-class
feasts land on consecutive admissible days in the one round they
collide, in band order.
Also folds in Task 5's review finding: year_bounds clamps y to [1582,
9999] once, up front, rather than guarding start and stop independently
(each guard only ever covered one of the two rite.year_start calls,
leaving year 999 and year 100000 each able to call it out of domain
through the other branch).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Precedence.resolution already tracked what happened to every losing
candidate -- commemorated, deferred, or omitted with a reason -- but
Liturgical_day.t had nowhere for the deferred and omitted buckets to land,
so Calendar dropped them at the door. Task 12's no-celebration-lost
invariant needs to read that accounting off the day result itself, not
re-resolve every day to reconstruct it, so a reason recorded nowhere is not
recorded.
Add Liturgical_day.omitted : ('r Celebration.t * string) list, after
transferred_out and before citations. Calendar.resolve_day now folds
resolution.omitted (Precedence's own native omissions, reasons intact) and
resolution.deferred (mapped to "deferred: transfer placement not yet
implemented (Task 6)") into it.
Adds a full-day accounting test against the whole Calendar pipeline: four
colliding sanctoral entries plus the day's feria, checked as a slug set
(matching test_precedence.ml's own nothing-silently-lost test) so a
candidate silently dropped or duplicated into two buckets would fail it,
plus an identity check that the deferred and admission-limit reasons don't
get swapped.
|
| |
|
|
|
|
|
|
|
|
|
| |
Transfers make per-date resolution impossible to do correctly: resolving 25
March can push a feast onto 26 March, and RG 97-98 has coinciding I-class
feasts transfer in table order, which needs global knowledge. So year computes
a whole liturgical year in one pass and day indexes into it. Pure, no cache, no
mutable state.
This commit resolves each day but does not yet place deferred transfers; they
are recorded with a reason. Task 6 adds the placement pass.
|
| |
|
|
|
|
|
|
|
|
|
| |
Validate took four loose arguments that had to come from the same rite with
nothing enforcing it, and Calendar is about to add more. Bundling makes a
mismatched assembly unrepresentable through the normal path.
season_runs replaces the hardcoded assumption that every season occupies exactly
one unbroken run. That holds for the 1962 rite but is false for the modern
form's Ordinary Time, which is one season in two runs -- as written the check
would have reported a false failure every year for the second rite.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The accounting test only checked bucket lengths, which a mutant satisfies by
duplicating a candidate across two buckets while dropping another entirely.
Replace it with a sorted slug-set comparison (Alcotest.slist), which a
duplicate-or-missing slug both fail.
Add two cases the brief's three properties call for but nothing exercised:
input-order independence (permuting the sanctoral list must not change the
outcome) and a temporal-only day (empty sanctoral list), the case the
temporal/sanctoral split exists to make safe.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Three rite-supplied functions, not one: band (who wins, RG 91), disposition
(what happens to the loser, RG 92-95) and admit (how many commemorations are
admitted, RG 111). The loser's fate depends on the loser's own rank, so
conflating them would resist extension.
resolve takes the temporal candidate separately from the sanctoral list, which
makes it total by construction. Every candidate lands in exactly one of
observed, commemorations, deferred or omitted -- nothing is dropped silently,
which is what makes the no-celebration-lost invariant checkable.
|
| |
|
|
|
|
|
| |
The 1960 reform reduced many feasts to a bare commemoration. They keep a rank,
because RG 111 orders admitted commemorations by dignity, but they can never be
the observed day. Modelled as a separate status rather than a fifth rank: RG 8
fixes the classes at four.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
week is "" for roughly 30 days a year (any day outside a numbered
week). Printed as-is among seven space-separated fields, that collapses
two fields into a double space, so naive field-position parsing (e.g.
awk '{print $4}') silently reads the wrong column on those days -- a
real defect for a project whose stated design is Unix-composable CLIs.
Emit "-" instead, so every line always has exactly seven
single-space-separated fields.
Record.headers/to_row exist for exactly this kind of column-safe
encoding but use a different column set and order than this CLI's
documented date/weekday/season/week/slug/rank/colour format (design
spec §6); adopting them here would be a larger, unrequested format
change, so this keeps the CLI's own field list and only substitutes
the empty value.
The cram test's pinned lines are re-verified against liturgical
anchors before promoting, not blind-promoted: 1 Jan 2026 (Circumcision,
class-1, white) and 2-3 Jan (ordinary class-4 Christmastide ferias,
the finding-1 slug fix already re-pinned separately) are correct;
2026-04-05 remains the sole Easter Sunday line, Paschaltide week 1,
class-1, white, matching the independently-computed Easter anchors
printed by `colitur easter 2026` immediately above in the same file.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Validate.run ~year:9999 raised (year_start (year + 1) asked year_start
for civil year 10000, out of the kernel's 1583..9999 domain), even
though 9999 is itself in range and kernel computation must never raise
on in-range input; ~year:9998 already returned zero failures. run now
clamps its scan to 31 December 9999 instead of computing year_start
(year + 1) when year is the domain maximum, and validates the
resulting truncated final liturgical year rather than not being able
to run it at all.
The design spec's validation §5 lists eight checks; only five were
implemented (coverage, seasons, weeks, weekday, closure). The two
missing were a real gap, not just a documentation slip:
- §5.7 anchor agreement. All of an EF year's Easter-derived and fixed
named days were pinned only by point assertions for 2026. run now
takes an ~anchors:(int -> (string * Date.t) list) parameter -- the
rite's own independent restatement of those dates, paired with the
slug each should carry, not derived from temporal itself -- and
checks that temporal agrees on every one of them. Temporal_ef.anchors
supplies EF's list. Kept rite-agnostic: the anchor list comes from
the rite argument, not the kernel.
- §5.8 determinism. run now calls temporal a second time for every
date and checks the result is structurally equal to the first.
Also, finding 8: the rank/season closure checks compare vocab entries
via their _to_string images, which is only sound if those images are
injective. run now checks List.map rank_to_string ranks and
List.map season_to_string seasons for duplicates up front and reports
a "vocab" failure if either collapses two distinct values to the same
string, rather than relying on that injectivity unasserted.
Test-quality fixes to the existing synthetic fixture, found while
adding coverage for the above: the fixture's own comment claimed its
mutation target (2026-03-15) was "not a Sunday" and "sits safely
mid-run" -- it is a Sunday, which made the coverage/week mutations
cascade further than documented even though the assertions still
target specific check labels. Moved to a genuine mid-week day
(2026-03-17) and the comment corrected. extreme_years's own test
required only "found at least one" of the two Easter-extreme years;
tightened to require both, since both genuinely exist in 1583..2500.
Covering tests: test_year_9999_does_not_raise (would error under the
old code; the fix is pinned by calling run 9999 directly with no try,
plus asserting the truncated year is reported via an ordinary
"seasons" failure, not silently or via coverage); anchor-clean and
anchor-fires cases on the synthetic rite; a determinism-fires case
using a target date whose temporal alternates what it returns across
successive calls; two vocab-injectivity-fires cases (collapsed rank
strings, collapsed season strings).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Two correctness bugs in Temporal_ef, both only visible across many
years, not a single point assertion:
1. Duplicate slugs within one liturgical year. Christmastide has no
numbered weeks, so the ferial fallback's <season>-<week>-<weekday>
scheme collapsed every feria in it to literal week "0". Since
colitur's Christmastide runs 25 Dec - 13 Jan (RG 72-73, a deliberate
divergence from lectio), the same weekday recurs across that span
and the keys collided (e.g. ef-christmas-0-saturday on three
different dates). Fixed with a dedicated christmastide_feria_slug
that splits the span into four sub-stretches, prioritising
compatibility with lectio's own keys wherever lectio has one:
- 26-28 Dec keeps lectio's existing ef-christmas-0-<weekday>.
- 2-5 Jan becomes ef-christmas-1-<weekday> -- lectio collapses
this indistinguishably into the same key as the stretch above,
so there is nothing to preserve; a colitur-only lectionary gap.
- 7-13 Jan splits at the actual first-Sunday-after-Epiphany
origin: on/after it, this is genuinely week 1 of Time after
Epiphany and takes lectio's own ef-time-after-epiphany-1-<weekday>
(which also can't collide with that season's own later week-1
ferias, since it's the same computation). Before it (0-6 days,
whenever Epiphany doesn't fall on a Saturday), a first attempt
at labelling this "week 1" too, matching a literal reading of
lectio's behaviour, was verified empirically (a throwaway sweep
of the full 1583..9998 domain) to reproduce duplicates in most
years -- so this remainder is its own ef-christmas-2-<weekday>,
a further colitur-only gap.
Verified with the same throwaway sweep: zero duplicate slugs across
1583..9998 after the fix.
2. named's week field was set by hand on some branches (Passion/Palm
Sunday, Easter, Low Sunday, Pentecost and its Vigil, Christ the
King) and left at None on others (Ascension and its Vigil, Corpus
Christi, Sacred Heart) even though all of them sit inside a
numbered season run. named no longer carries a week at all --
temporal now calls week itself for every day, named or not, so "a
named day inside a run carries that run's week" holds by
construction rather than by remembering to set it on each branch.
temporal.mli's week field doc is reworded to state the actual rule.
Covering tests: point assertions for all four Christmastide
sub-stretches; a QCheck property scanning random years for any
duplicate slug within a liturgical year (excluding the deliberate
resumed-Sunday reuse); point assertions for the four previously-None
days now carrying their week (first to fail without the fix: Ascension
Vigil, expected Some 6, got None); a QCheck property asserting
temporal's week equals week for every day of the year, replacing a
prior property whose fallback made it structurally incapable of
detecting an omitted week.
Also: cite RG 91 e18 for the September/Advent Ember days matching
lectio and flag the Lent/Whitsun Ember and Rogation slugs as
colitur-only lectionary gaps inline, matching the existing
Nativity-vigil/octave-day convention; cite RG 117/123/127/128/131 for
season colours and Gaudete/Laetare rose; drop the unreachable
Passiontide arm from the Sunday-rank match (Passiontide has only two
Sundays and both are already named above, so no Passiontide Sunday
ever reaches that fallback).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Layer.load narrowed its catch to Sexplib0.Sexp_conv_error.Of_sexp_error,
but rank_of_sexp is caller-supplied and may raise anything -- e.g. a
hand-written rank parser that calls invalid_arg. Overlay.load already
catches every exception from the equivalent call; mirror that here so
layer.mli's "never as an exception" promise actually holds.
Date_spec.t derived its sexp converters with plain ppx_sexp_conv, unlike
Slug and Lang, which hand-write validating parsers specifically so
malformed data is rejected at load. (Fixed (month 13) (day 40)) used to
deserialise cleanly into a spec that simply never resolves -- a saint
quietly vanishing with no diagnostic. t_of_sexp now re-runs the value
through the existing fixed validator, the same shape Slug and Lang
already use.
Covering tests: a Layer.load case where rank_of_sexp raises
Invalid_argument instead of Of_sexp_error (would have escaped
uncaught before this fix); two Date_spec.t_of_sexp cases (month 13,
31 April) that must raise Of_sexp_error rather than silently
constructing an unresolvable spec.
|
| |
|
|
|
|
|
| |
One line per day -- date, weekday, season, week, slug, rank, colour -- built
through Record, the canonical output view. Until Plan 3 brings the differential
and oracle layers, this dump is how a human checks EF temporal against
missalemeum by eye.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Task 14 review, findings 1 and 2.
Finding 1: the only tests against Validate exercised the clean path against
real EF data, so the evidence that each check can actually fire lived in a
scratch mutation probe that was never committed. A future edit that quietly
weakened a check would leave the suite green, since a weaker check only
makes more inputs pass. Added a small synthetic two-season, two-rank rite
fixture in test_validate.ml -- not EF -- letting each test violate exactly
one invariant directly: a temporal that raises (coverage), a season that
recurs (seasons), a week that decreases mid-run (week), a weekday that
disagrees with Date.weekday (weekday), a rank absent from the declared vocab
(rank), and a colour outside Colour.all (colour, via a same-representation
Obj.magic value, safe here because the check compares by structural equality
rather than pattern match). A clean-baseline test confirms the fixture itself
reports zero failures before any mutation is applied.
Each new test was verified non-vacuous by temporarily weakening its
corresponding check in validate.ml, confirming the matching test fails, then
reverting -- the same trap one level up, checked explicitly rather than
assumed.
Finding 2: removed the slug well-formedness check. Slug.t is a private string
validated on every construction path, and to_string is the identity, so
round-tripping an existing Slug.t can never fail -- the check was structurally
incapable of firing. Folded the explanation into the comment block that
already covers why slug uniqueness isn't checked, since it's the same kind of
fact: a property the type system delivers, not one Validate needs to assert.
The colour check stays: unlike slug, Colour.all is a hand-maintained list
that can drift from the type, so it is only practically (not structurally)
tautological, the same class as the rank closure check.
|
| |
|
|
|
|
|
|
|
|
|
| |
Coverage, season contiguity and completeness, Sunday-aligned week numbering,
slug well-formedness, weekday agreement and vocabulary closure.
Checks run over a liturgical year rather than a civil one, since Christmastide
straddles January and would otherwise appear to recur.
Run against EF temporal for landmark years, both Easter extremes, and 200
random years across 1583..9998 -- the property layer is how confidence reaches
past the oracle horizon.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
named hardcoded week = None for ef-pentecost-vigil, ef-pentecost and
ef-christ-the-king, unlike the five sibling named Sundays (Easter, Low
Sunday, Passion Sunday, Palm Sunday, Trinity), which already carry the
explicit week number that week computes for the same date. Because these
three sit inside a season run rather than at a run boundary, the omission let
the week number jump on the following Monday without an intervening Sunday
marking the change -- breaking week continuity in every single year in the
domain.
This was found by the new Validate invariant harness (Task 14), which
flagged exactly two failures in every one of the 8416 years 1583..9998, both
on the Monday after one of these days. Season contiguity itself was clean
across the whole domain; only week numbering was affected.
Pentecost and its Vigil are fixed Easter offsets, so they take the literal
values (7 and 8) that week already computes for them. Christ the King's date
varies by year, so the same expression week evaluates is inlined instead,
since week is defined later in the file and cannot be called from named.
Added a general property test, prop_named_week_agrees_with_week, checking
across 200 random years that wherever named gives an explicit week, it
matches what week independently computes for that date -- covering all
sibling entries, not just the three fixed here, so the same drift cannot
silently reappear in a future addition to named.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Ferial ranks per RG 91: I class in Holy Week and the privileged octaves
(entries 7, 10), II class for Advent 17-23 and the Ember days (entry 18), III
class for Advent to 16 December and Lent/Passiontide (entries 22, 25), IV class
per annum (entry 28). Two further divergences from lectio, both from the table:
Advent 17-23 are II class, and the Lenten Ember days are II class rather than
ordinary Lenten ferias. Rogations (RG 80/87) exist at all, which lectio does
not compute. Gaudete and Laetare are rose.
Two corrections against the primary source beyond the task brief's draft.
Entry 7 (verified against the Latin: "feria IV cinerum et II, III et IV
Hebdomadae sanctae") covers only Ash Wednesday and Monday-Wednesday of Holy
Week; Thursday-Saturday are the Sacred Triduum, entry 2, cited separately.
Rogation Monday and Tuesday keep the ordinary ferial rank of their season
(RG 88: "de Litaniis minoribus nihil fit in Officio") rather than a fixed
class, since no RG 91 entry ranks them specially.
A module-type constraint proves Temporal_ef satisfies the kernel RITE contract.
|
| |
|
|
|
|
|
|
|
| |
Task 12 review: add a permanent regression test for floor_div's negative
branch (Ash Wednesday, 4 days before the Lent I origin -- the only date in
the system where it fires), and stop sunday_slug from independently
recomputing the Pentecost-relative week number that week already computes.
A cross-check test pins the two together before the refactor and continues
to guard against future duplication.
|
| |
|
|
|
|
|
|
| |
Week origins are Sundays, so week numbers are constant Sunday-to-Saturday; the
origin is season-specific, which is why Validate must not assume a run starts
at 1. When Easter is early the surplus Sundays after Pentecost resume the
Sundays after Epiphany that Septuagesima cut short, and the last Sunday before
Advent always keeps the 24th (Last) Mass.
|
| |
|
|
|
|
|
|
|
|
| |
The I-class feasts of the Lord (RG 91 entries 1-3), the Nativity, Ascension
and Pentecost vigils (RG 28-34), and the days within the Octave of the
Nativity (RG 64-70). Christ the King is the last Sunday of October per the
1960 calendar, not the OF's last Sunday before Advent.
Rogation Wednesday coincides with the Ascension vigil; with no precedence
framework until Plan 3, temporal emits the higher-ranked vigil.
|
| |
|
|
|
|
|
|
| |
Christmas Time runs to 13 January inclusive (RG 72-73) and time after Epiphany
opens on 14 January (RG 77), diverging from lectio, which starts
time-after-epiphany at 6 January. Holy Saturday stays in Passiontide: the
Easter Vigil is a night Mass, and a per-day calendar assigns a date by the
season governing its day-hours.
|
| |
|
|
|
|
|
| |
Eight seasons in canonical liturgical-year order, each carrying its RG
citation, and the four classes. season_slug_word is deliberately distinct from
season_to_string: slugs are lectionary keys adopted verbatim from lectio,
which calls Paschaltide 'easter' and Christmastide 'christmas'.
|
| |
|
|
|
|
|
|
|
| |
Make alignment impossible: both headers and to_row are derived from a single
list of (name, extractor) pairs, so reordering one without the other is
impossible. CSV schema cannot silently mislabel columns.
Add a meaningful test that pins the header names in order, not just length.
This catches column reordering regressions.
|
| |
|
|
|
|
|
| |
The boundary where rite-parametric types stop. CSV, JSON and the template
engine all render from this one schema, so they never see a type variable.
headers/to_row cover the scalar columns; names and citations are
variable-arity and wait for the richer encoders in Plan 5.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Five review findings, all in the test suite the task designates as the
algebra's specification:
- test_diagnostics's headline assertion was vacuous: it checked slug
presence after an add-over-existing, but hilary was already present
before the Add ran, so a broken implementation that recorded the
diagnostic and skipped applying the entry would still pass. Add now
carries a rank that genuinely differs from base()'s default, and the
assertion checks the field actually changed.
- Set_subject, Set_name, Remove_name, Set_citation and Remove_citation
were implemented but never exercised. Added coverage for all five,
including the case that matters most for Set_citation: setting a part
that already exists must replace it, not append a duplicate.
- Replace had zero coverage. Added tests for both its existing-slug path
(entry swapped, no diagnostic) and its absent-slug path (diagnostic
recorded, entry still added).
- merge's diagnostic accumulation was untested; test_merge_order discarded
the diagnostics it got back. Added a test asserting two overlays' each
producing one diagnostic come back in application order.
Extended the cel/entry test helpers with optional subject/names/citations
so the new tests can build fixtures with those fields populated.
|
| |
|
|
|
|
|
| |
add/suppress/replace/field-edit folded in order, last writer wins per field,
empty the identity. A directive naming an unknown slug, or adding one that
already exists, yields a diagnostic rather than silence or a hard failure:
overlays must survive a shifted base while still surfacing authoring errors.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
load previously had zero coverage despite being the module's only fallible-
I/O function and the one place the brief's literal code would have let an
exception escape on this sexplib version (Failure instead of Parse_error).
Add cases for a missing file, a malformed/unterminated sexp, an invalid slug
embedded in an otherwise well-formed layer, and a success round-trip.
Also add a same-date index case with three entries inserted out of slug
order, pinning that the by-date index accumulates rather than overwrites and
returns buckets in canonical order regardless of insertion order -- the
previous two-entry fixture only ever touched bucket sizes 0 and 1.
|
