colitur.git/lib/kernel, branch v1.2.0

fix(of): anticipate St Joseph onto 18 March per Normae n. 56(f)

2026-08-26T21:40:17Z

W1: when 19 March (St Joseph) falls on Palm Sunday, colitur sent him forward to Easter+9 instead of anticipating the solemnity backward to 18 March, the one rule in the whole Normae transfer machinery that names an earlier date. Live in 16 of 1583-2100 (1595, 1606, 1617, 1690, 1758, 1769, 1780, 1815, 1826, 1837, 1967, 1978, 1989, 2062, 2073, 2084). Blocked on a kernel contract: Rite.t.transfer_target required its result be strictly later than the impeded date, an obligation argued purely from the EF's RG 96. Traced against Calendar's actual placement pass rather than assumed unsafe: the round loop re-resolves the whole year fresh every round, and its two safety nets (~start/~stop; the round-count guard) both test the target's own value with no comparison against origin, so a bounded backward jump converges the same way a bounded forward one does. The one requirement that stays load-bearing is target <> origin: Precedence.resolve never deduplicates by slug, so a candidate returning its own argument unchanged would collide with itself in the same round's contest. Relaxed the obligation accordingly (rite.mli) and added the fourth transfer_target branch in Precedence_of, ahead of the existing three (Precedence_of.transfer_target, Rule 0). calendar.ml needed no logic change, only two comments that asserted "every genuine transfer moves forward" as a blanket fact. Unpinned the two tests that asserted the old wrong (forward) behaviour and replaced them with assertions of the correct one, plus a new direct unit test of the backward branch across two independent years. EF output is untouched: data/ef/ and lib/rites/rite_ef/ carry no diff, and colitur day/readings for 2026, 1583 and 9999 are byte-identical against a build of the prior commit. OF's own 2005-2050 window carries none of the 16 affected years and shows zero unexpected diffs; the full 1583-9999 domain still reaches a fixed point under the exhaustive sweep.

fix(citation): make the well-formed reading-part set rite-supplied

2026-08-26T12:28:50Z

I7 (final-review.md, 2026-08-25-colitur-of-phases-3-5): OF Sundays and solemnities shipped without the Second reading, and the lectionary data header called the First reading an "Epistle" -- EF vocabulary. The sole blocker was Validate's own kernel-hardcoded assertion that a well-formed day's sorted citation parts are exactly [First; Gospel]; Citation.part already carried Second. Rite.t gains citation_shapes : Citation.part list list, the fourth EF-shaped kernel misfit (after Preface.t, Mass_formulary.source, transfer_target's strictly-later contract) and the first fixed rather than only documented. EF supplies its one shape unchanged; OF supplies two, [First; Gospel] on a feria/feast/memorial and [First; Second; Gospel] on a Sunday or solemnity (OLM 1981 Praenotanda n. 66.1 vs n. 69.1, n. 84(b)/(c), page-image verified against olm-1981.pdf). Validate's "citations" check now matches against the rite's own declared shapes rather than a single literal pattern, preserving all three invariants it always enforced: zero or one complete shape, never partial; no part outside what the rite declares; no part repeated. bootstrap_lectionary_of.ml's cite now extracts a Second reading from Sunday-cycle ini sections only, correctly excluding the one stray `second` field lectio's snapshot carries on a weekday-cycle section (a content-verified scraping artifact). The hand-authored Christmas Day entry gains its own Second reading from the same OLM passage already cited for its First/Gospel. Regenerated data/of/lectionary.sexp: 771 entries unchanged, citation fields 1542 -> 1725 (+183, all Second readings), SHA-256 re-pinned, header wording fixed and a new coverage bullet added. Responsorial Psalm is deliberately still not emitted -- OLM 1981 Praenotanda n. 71 itself lists it apart from the readings proper as a chant with its own refrain, and the finding driving this change names only the missing Second reading. EF output verified byte-identical to the branch start (empty lib/rites/rite_ef + data/ef diff since v1.0.0, plus a direct binary comparison of `day`/`readings` for 2026, 1583 and 9999 against a worktree build of the pre-change commit). Tests: three new fixtures in test_validate.ml prove the generalisation holds both directions (a multi-shape rite accepts either of its declared shapes; a rite that never declared a shape still rejects it). test_citation_coverage_of.ml's pinned residual grows 41 -> 49, all new entries the same already-understood chapter-crossing shape.

feat(ef): the Mass preface, RG 482-499

2026-08-22T21:24:12Z

Item 2 of Phase 3 (celebrant-rubrics-phase1), the bulk of this phase and the last EF Mass rubric this project scoped: which preface is said, deferring to seasonal/proper-title rules for the fourteen named prefaces, then RG 498's Common residual. RG 482 gives the resolution chain: the Mass's own proper preface, failing that the seasonal one, failing that Common. Read literally, RG 484-497 look like fourteen separate rules, but each numbered rubric's own propria/de-Tempore pair produces the SAME preface identity either way, so the whole chain collapses into one priority- ordered decision: title/mystery triggers (Holy Cross, Sacred Heart, Christ the King, Trinity, St Joseph, BVM, the Nativity octave, the Apostles, Epiphany), each independent of season, then six seasonal windows (Nativity, Epiphany, Lent, Holy Cross/Passiontide, Easter, Ascension, Holy Spirit, Trinity-for-ordinary-Sundays), then Common. The "one genuinely unproven piece" the design spec worried about -- per-feast proper prefaces extracted from the Missal's propers at scale -- turned out not to be a large-scale extraction problem at all: every one of the fourteen propers is a closed, small, subject/slug-keyed trigger (mirroring Precedence_ef.band's own RG 91 table), not thousands of individual saints' pages. The one genuine open question (RG 488, the Chrism Mass) is N/A: this engine resolves one Mass per civil day and has no separate Chrism-Mass dimension. Preface is a new kernel type (lib/kernel/preface.ml[i]), not an EF-specific one, the same placement as Colour/Subject/Mass_formulary: Liturgical_day.t is parameterised only over season/rank, so any field it carries generically must live in the kernel even though only EF constructs a value of it today. Wired exactly as creed/gloria were (Rite.t.preface, Calendar.ml, Rite_ef.context), but Preface.t option, not a bare bool: unlike creed/gloria, a preface is said only at a Mass, and Good Friday (1955-restored Holy Week) resolves an observed celebration but has no Mass at all -- None is the honest answer there, and also the neutral value an unimplemented rite returns. Two priority-order findings only the oracle settled, not derivable from the Latin text in isolation, both cross-checked against 358 individually classified entries in the FIUV Ordo's own praef column (test/fixtures/fiuv-ordo-2025-2026.sexp, already captured, wired up here as a new comparison axis): RG 484(b)'s own "except Masses with a proper of the divine mysteries or Persons" is narrower than every other window's implicit exception (an Apostle inside the Nativity octave is overridden to Nativity; outside it, keeps his own preface even inside another window); and RG 495's "et votivis" half is live for the one office this engine models without a votive-Mass dimension (the Saturday Office of the BVM), which also proved a vigil is not a "festum" for this purpose (the Assumption's own vigil takes Common, not BVM, correcting an initial reuse of Precedence_ef.marian_slugs that had no reason to make that distinction for its own, different rubric). A third, RG16(a)-shaped fix landed the same way: RG 494(b)'s own Trinity grant must read the day's TEMPORAL season, not the celebration that actually won it, or a Class1 feast with no preference of its own (All Saints) wrongly falls to Common on a Sunday it merely commemorates. colitur rubrics gains a sixth TAB-separated column. Domain-wide 1583-9999 exhaustive sweep confirms every Christmastide day resolves Nativity/Epiphany/Bvm, every Paschaltide day one of its own three windows or a season-independent title, every Lent day Lent or a title, and Passiontide legitimately produces Easter exactly once a year (the Vigil Mass) -- 8416 of the domain's 8417 years, the one short year being the domain's own upper boundary (the liturgical year opening in 9999 cannot construct dates in year 10000, a pre-existing edge this sweep re-confirms rather than a new one). day/readings verified byte-identical to the branch's own state before this phase (commit 9c96e0a) across a 455-year sample spanning the whole domain -- not literally to the v0.10.1 tag, which 22 earlier commits on this same branch (Phase 1/2, the bissextile shift, the Rogation colour fix) had already moved past before this phase began.

feat(ef): RG 111(a), the sung-Mass commemoration cap

2026-08-22T20:40:07Z

Item 1 of Phase 3 (celebrant-rubrics-phase1): the "at Low Mass" commemoration-placement rule the design spec recorded as unread. It is not a new rule -- it is the sung/low axis of RG 111, which colitur already implements. RG 111(a) (LT.txt, "Ratio admittendi commemorationes"): a liturgical day of the first class, AND any non-conventual sung Mass regardless of the day's own class, admits at most one commemoration, and only if it is privileged. (b)/(c)/(d), the same rubric's remaining clauses, give the wider caps colitur's admit already computes -- which is exactly the LOW MASS answer. Exposed as Precedence.sung_mass_commemorations, a pure derivation over the existing Low-Mass admitted list (filter to Privileged, keep the first), not a new stored field on Liturgical_day.t: the input list is already validated and privilege-tagged, so a second field would only create a second place for the same fact to drift out of sync with the first, with no new information gained. Liturgical_day.t.commemorations is now documented as the Low Mass set explicitly, removing the ambiguity its .mli previously left unstated. Tested against synthetic Low-Mass sets (none/one/two privileged, already-first, empty) and two real calendar days resolved through the normal Cal.day pipeline: 2026-08-14 (Vigil of the Assumption, an ordinary-only commemoration, dropped at Sung Mass) and 2026-04-25 (the Major Litanies, RG 80/109(f), privileged, kept at both Masses).

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

2026-08-22T15:56:42Z

Phase 2 of celebrant-rubrics-phase1: colitur rubrics gains a fifth column, whether the Gloria is said. Follows the Creed's own seam exactly -- Rite.t.gloria, Liturgical_day.t.gloria, wired through calendar.ml the same way. RG 431(a)/432(a) defer the Gloria to the Breviary's own Te Deum rule (nn. 237-238), so te_deum is implemented as its own named predicate, cited clause by clause, not collapsed into a colour heuristic. 431(c) (Holy Thursday, the Easter Vigil Mass) and 432(b)/(d) (violet; a Requiem) are independent overrides checked ahead of the Te Deum-derived answer. Every clause this engine has no dimension to model (votive Mass classes, the wider n.302 "Missa festiva" categories) is stated as N/A with its own reasoning, not silently dropped. Validated against the FIUV universal Ordo (Gloria and Te Deum) and all three LMS editions (Gloria). A first pass over-trusted a clean-looking 15-for-15 FIUV contradiction of 237(b)'s own Septuagesima exception and replaced it with a blanket "every Sunday" rule; the evidence was itself corrupted -- the FIUV extractor recognised only one of the source's two Te Deum negations ("non dicitur", not "sine"), so every "sine Te Deum" Sunday read wrongly true. Fixed in tools/extract_fiuv_ordo.ml, fixture re-extracted, and the literal 237(b) reading restored once the corrected data confirmed it. A second bug surfaced alongside it (Palm/ Passion Sunday wrongly reading true via Temporal_ef.named's own table membership, then Christ the King wrongly reading false from an over-broad fix) is closed with an explicit two-slug exclusion. Domain-wide 1583-9999: every violet or Rose day is gloria=false except the Easter Vigil (RG 431(c) lex specialis), every Requiem is gloria=false, both measured exhaustively, zero exceptions. Mutation- proved: disabling 431(c) reddens 8 tests including all four oracle comparisons; disabling 238(c)'s feria-I-classis exclusion reddens exactly the dedicated Ash Wednesday unit test, a genuine blind spot in both oracle layers, reported rather than hidden. Two open, cited findings, neither fixed here (out of this task's "follow creed's exact seam" scope): a privileged Lenten/Passiontide feria carrying one commemoration reads Gloria=true in the LMS Ordo but Te-Deum=true/Gloria=false in FIUV -- the two oracles disagree with each other, not merely with colitur (data/ef/expected-divergences-lms.sexp L5, expected-divergences-fiuv.sexp F3); and a pre-existing, uncited Colour.Violet bug on Rogation Monday/Tuesday in Temporal_ef.temporal, surfaced by this comparison but root-caused as a separate defect (L6). day/readings verified byte-identical against a build from the branch tip before this task (v0.10.1's own tag predates an already-landed bissextile fix that legitimately changed both, so it is not the right baseline). 671 tests green (dune test); 678 with the exhaustive sweep (COLITUR_EXHAUSTIVE_SWEEP=1 dune test --force, ~104s).

feat(ef): implement the calendarium's bissextile February shift

2026-08-22T14:40:24Z

The Missale Romanum's calendarium footnote for February (LT.txt:5011-5014, scan-corroborated scan2.txt:3050-3058): in a leap year the sixth kalends of March (24 February) is doubled rather than a 29th day appended, so St Matthias moves 24->25 February and St Gabriel of Our Lady of Sorrows 27->28, with 24 February itself carrying no fixed office that year. data/ef/expected-divergences-lms.sexp entry L3 tracked this as an open gap. Implemented as the rubric's own general mechanism (every fixed entry from 24 through 28 February shifts one civil day later in a leap year), not as "move these two named saints": the two readings are indistinguishable on shipped data (nothing else is fixed in that window) and diverge only for a future --overlay entry in the same range, which the mechanism reading covers correctly and a two-saints special case would not. This project has already been bitten once (RG 16(a)) by a rule implemented against shipped data's coincidental shape rather than the rubric itself. Kernel stays rite-agnostic: Rite.t gains a fixed_key field (Date.t -> (int * int) option), the same seam easter already establishes, read only by Layer.on_date's FIXED half -- the movable half (Easter_offset/Nth_weekday) is untouched. Default is the identity mapping, an optional argument on on_date, so every existing caller and every rite that supplies nothing is byte-identical to before this field existed. Date.is_leap is exposed in date.mli (already existed in date.ml) so the rite reads the kernel's own single-sourced leap-year definition. The EF implementation lives in Rite_ef.Temporal_ef.bissextile_fixed_key, fully cited, wired into Rite_ef.context. Blast radius, measured over the full 1583-9999 domain (Calendar.year run twice per civil year, real fixed_key vs identity, every day diffed): all 2,041 leap years in the domain show a visible change for each saint; 6,983 individual liturgical days change total, zero unclassified, across four shapes (Matthias vacates 24 Feb in 1,803 years / occupies 25 Feb in 1,811; Gabriel vacates 27 Feb in 1,699 / occupies 28 Feb in 1,670). L3's own prior estimate (1,650 years, observed-outright only) is confirmed close on the same definition (1,677 measured); the broader observed-or- commemorated count is higher, not lower, showing the narrow estimate undercounted visible impact rather than overcounting it. The lectio differential (2005-2050 fixture) gains a new cited entry, C40 in data/ef/expected-divergences.sexp: lectio implements no such shift, so every leap year in its window now diverges on the two shifted days (21 rows, not the naive 44 -- the comparator never compares commemorations, so a side where the shifted saint has zero comparator-visible footprint on both engines produces no row). The LMS Ordo's 2023-2024 fixture -- the window L3 was originally found in -- now matches on the Creed comparison; its formulary-override bucket count is corrected 182->181 (24 February leaves the counted Proper population, becoming Own_slug-sourced). The missalemeum oracle fixtures (2026-2027, 2038, 2035) cover no leap year and are unaffected. L3 is closed: converted from an active allow-list record to a prose closure paragraph (L1's own precedent), citation preserved verbatim, fix and measured blast radius recorded. The id-list assertion narrows to [L4] alone, and the now-dead "2024-02-24" -> "L3" date mapping is removed. Six golden values pinned across four new test cases: Matthias in a leap year (both the vacated and occupied sides) and a common-year control; Gabriel likewise, deliberately choosing a leap year where he is admitted only as an ordinary commemoration rather than winning outright, a different shape from Matthias's. Mutation-tested: forcing fixed_key back to identity reddens exactly five test cases -- the differential's Layer C count-pin (C40 drops to 0 actual vs 21 expected), the LMS Ordo's Creed and formulary-bucket checks, and both new golden pins -- confirming the tests actually exercise the fix. Reverted; dune test, the exhaustive sweep, and make check are all green.

fix(kernel): Mass_formulary.t.said is honestly optional -- was false for Votive

2026-08-22T11:48:08Z

The .mli promised said is "the slug whose Mass is said". For Votive (RG 78/309(a), the Saturday votive Mass of Our Lady) it was set to the day's own ferial slug -- whose Mass is exactly the one NOT said. A consumer joining rubrics to readings on that slug would silently get the wrong Mass: 2026-01-03 reports ef-christmas-1-saturday, which has zero entries in data/ef/lectionary.sexp, because the citations actually come from bvm_saturday_citations, a season-keyed function with no slug of its own anywhere in the shipped data. Chose the type-honest fix over the interim documentation one: said is now Slug.t option, None exactly for Votive, because there is genuinely no slug in the shipped data this field could report for that one source. Adding real ids for the five seasonal BVM Masses (the reviewer's first option) is out of scope -- a data restructuring this round explicitly does not carry. Threading the office slug through a second field was considered and rejected as redundant: the day's own office is already available on the same Liturgical_day.t via observed.slug, which every caller already has in scope regardless of via, so said does not need to duplicate it. colitur rubrics stays byte-identical: rubrics_line already has d.observed in scope and falls back to its slug when said is None, printing the exact value it always printed for a Votive row (verified directly, diffed against pre-fix output across four years). colitur day/readings are unaffected (neither reads Mass_formulary at all). colitur emit --format sexp's pretty-printed line count for 2027 moved 9011 -> 9025: every day's formulary record widened by said's own extra option wrapping, and to_string_hum wraps by column width. Cosmetic only, diffed line by line to confirm every change is this shape or a consequent wrap shift; recorded in test/cli.t alongside the 476(f) note it now sits next to.

feat(ef): the Creed, RG 475-476

2026-08-22T10:02:01Z

Whether the Creed is said at Mass. New Rubrics_ef.creed, reached through a new Rite.t.creed field and a new Liturgical_day.t.creed bool (false, not an option, for a rite that has not implemented the rule); colitur rubrics gains a fourth TAB-separated column. 475(a) reads the TEMPORAL cycle's own weekday, not the observed day: a Sunday impeded by a Feast of the Lord (RG 16a) still says the Creed. 475(d)'s three octaves (Nativity, Easter, Pentecost) are pure date/ Easter-offset windows, checked first, since a saint's feast winning the day inside one of them still says the Creed ("etiam in festis occurrentibus") -- St Stephen, 26 December, is the live witness. RG 23 classifies Ash Wednesday and every feria of Holy Week (including the Sacred Triduum) as feriae, not festa, however high their rank, so 475(b)/(c)'s "in festis" never reaches them; this single check subsumes 476(a)'s own naming of the Chrism/Lord's-Supper Mass and the Easter Vigil. RG 28-34 vigils are a third liturgical-day category, also excluded from "in festis" regardless of rank -- reuses the already- exported Precedence_ef.is_vigil rather than a new list. 475(e)'s Apostle/Evangelist natalicia list (creed_apostle_slugs) was derived by grepping data/ef/sanctoral.sexp directly and checking each candidate's own date against whether it is that saint's dies natalis -- not copied from any list supplied with the task. The Conversion of St Paul and the 30 June Commemoration of St Paul are excluded (neither is a natalicium); the Chair of St Peter and St Barnabas are included only because the clause names them explicitly, which is exactly why it has to: neither is a natalicium either. 475(c)'s BVM half reuses Precedence_ef.marian_slugs (newly exported) rather than the subject field alone: checked against the shipped data, almost every Marian sanctoral entry ships subject=Saint, not Bvm. man/colitur.1's rubrics section is updated to match the new column; test/cli.t repinned via dune promote for the same reason. Verified day/readings output byte-identical to v0.10.1 across the whole 1583-9999 domain (both binaries' concatenated day+readings output, 6,148,492 lines each, zero diff). Domain-wide: 882,996 days say the Creed, 2,191,250 do not; every one of the domain's 439,178 Sundays says it, zero exceptions.

fix(kernel,ef): a votive Mass is not the day's own -- add Mass_formulary.Votive

2026-08-21T20:54:13Z

Mass_formulary.source was missing a case for a Mass said IN PLACE of the day's own office's Mass while that office is itself kept unchanged -- RG 309(a) and RG 431(e) both classify the RG 78 Saturday Mass of Our Lady this way, in the Missal's own words, as a "Missa votiva IV classis... de B. Maria Virg.", not as the day's own office's Mass. The Latin Mass Society Ordo (docs/research/ordo/lms-ordo-2024-2025.pdf) witnesses it directly, printing that day as "V Mass of BVM". The BVM-Saturday branch in Lectionary_ef.readings was tagged Own_slug for lack of a better constructor when Task 2 landed, flagged there as a judgement call rather than a specified answer. That call was wrong: left as Own_slug, a future comparison against the LMS Ordo (a later task in this plan) would read every BVM Saturday as a manufactured divergence between colitur's "own" Mass and the Ordo's votive one. Retagged to Votive; said is unchanged (still the day's own, reused ferial, temporal slug) since the office itself is unaffected, only the Mass said for it. Added the new constructor's source_to_string case ("votive") and its own test row, and pinned the branch itself in test_lectionary_ef.ml's formulary cases at 1 August 2026, verified directly against the resolver rather than assumed. Both this session's own drifted pins that the earlier commit inherited from the task brief are unaffected by this change, and both are re-confirmed independently correct in this round: dune test and the exhaustive sweep are green, and colitur day stays byte-identical against the pre-fix-round binary.

feat(kernel,ef): the lectionary reports which Mass it said

2026-08-21T20:39:12Z

Rite.readings now returns (Mass_formulary.t option * Citation.t list) instead of a bare citation list, and Liturgical_day.t carries the result as a new formulary field. Validate holds a rite that resolves a formulary at all to resolving one on every day, the same discipline it already applies to citations; the EF lectionary chain resolves Some on every day of every year 1583-9999, confirmed by a direct sweep over 2005-2050 as well as through Validate itself. Plan Tasks 2 and 3 are merged into this one commit on the coordinator's own instruction: Rite.readings' signature and the field that consumes it are one atomic edit, and the intermediate state does not compile on its own. Each of the four lectionary steps now builds its own Mass_formulary.t at the point it decides, not by re-deriving it afterwards from the citations it returns: step 1 tags Proper with the observed slug, step 2 tags Own_slug with the day's own temporal slug, step 3 tags Preceding_sunday with the resumed Sunday's temporal slug, and step 4 tags Common with the Common's own id -- Commons.find now returns that id alongside its citations rather than discarding it, since it is only ever in scope at the point the assignment is looked up. The RG 309(a) Saturday votive Mass of Our Lady, which answers between steps 4 and 2 rather than as one of the four numbered steps, is tagged Own_slug too: Mass_formulary.source has no dedicated constructor for it, and its own guard only ever fires when the observed celebration already is the day's own (reused ferial) temporal slug, which is exactly what Own_slug documents. Recorded as a judgement call in the task report, not a specified answer. test/cli.t's `emit --format sexp` line count is repinned (8472 to 8881): that command serializes Liturgical_day.t whole, so the new field grows its output. `colitur day` itself is untouched -- verified byte-identical against the pre-change binary across 1583, 1900, 2026, 2038 and 9999.