1 files changed, 6 insertions, 0 deletions
diff --git a/CHANGELOG.md b/CHANGELOG.md
index fc2ac99..89b2281 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,12 @@
 
 One line per release. Versions are git tags (`vX.Y.Z`).
 
+## [1.0.0] - 2026-08-24
+- **colitur now prints what a real ordo prints for the Mass.** That is the whole of this release, and the scoping rule it was built to: nothing more, nothing less. `colitur rubrics <year>` gives, per day, which Mass formulary is said and how that was reached (`proper`/`own`/`preceding-sunday`/`common`/`votive`), the Gloria, the Creed and the preface; commemorations gained their Low-Mass/sung distinction. **The Creed** is RG 475-476, whose clause (a) reads the TEMPORAL cycle rather than the observed day — *"in qualibet dominica, etsi eius Officium alicui festo locum cedat"*, a Sunday impeded by a feast still says it — and whose clause (e) turns on *natalicia*, which is why the rubric must name the Chair of St Peter and St Barnabas explicitly and why the Conversion of St Paul is excluded; the Apostle list was derived from the shipped data and verified entry by entry rather than transcribed. **The Gloria** is RG 431-432, whose clause (a) defers a Mass question to the Breviary — *"quotiescumque ad Matutinum dictus est hymnus Te Deum"* — so the Te Deum is implemented as its own cited predicate from Breviary nn. 237-238 rather than collapsed into a colour heuristic. **The preface** is RG 482-499's chain, proper else seasonal else common, with RG 483 bounding it: no commemoration ever induces a proper preface. **Chants are deliberately absent** and always will be: a priest has the Missal open, and an ordo's job is what the book cannot tell you when you open it.
+- **Four rubrical defects, every one found by an external witness rather than by inspection.** *The Creed was said at Requiem Masses* — RG 476(f) forbids it, and All Souls had been wrong every year; closed via a `Colour.Black` proxy sound because black occurs exactly twice in the whole engine. *The bissextile shift was missing entirely* — the calendarium's own February footnote moves St Matthias to 25 February and St Gabriel to 28 February in a leap year, *"et bis dicitur sexto calendas"*; 2 041 leap years, 6 983 days, and it surfaced only because a witness year covering 2024 was added. *Rogation Monday and Tuesday were violet* — RG 119(b) makes Paschaltide white and RG 128 does not list Rogations at all; the violet belongs to the votive Rogation Mass, not the day's office, and fixing it cascaded correctly into the Gloria. *The ferias after the Ascension resumed the preceding Sunday's Mass* — the Missal's own marginal rubric repeats the Ascension's through the week; 16 350 days.
+- **The single-lineage problem is broken, not merely documented.** Layers 3 and 4 both descend from Divinum Officium, so for two years this project had one witness counted twice. It now has **three independent publishers over seven witness-years** — the FIUV universal Ordo (Latin, recording the Te Deum 262 times, and so a second witness on the Breviary clause whose own source is a single web transcription), three Latin Mass Society editions, and three from extraordinaryform.org. Preface agreement is 358/363, 358/363, 361/362, 395/395, and three LMS years at 396/396, 399/399, 398/398. Each source was characterised before use — probed on dates where the rubric predicts OPPOSITE answers — because an ordo prints a MENU of permitted Masses and comparing against the wrong item manufactures divergences: that trap produced 200 false colour findings, 12 false class findings, and one adjudication that had to be publicly withdrawn, all inside this release's own development.
+- Also: RG 33's own second-order consequences settled; RG 111(a)'s sung-Mass cap exposed as a derivation over data already validated rather than a new stored field; RG 497 shown to lack the *natalicium* restriction RG 475(e) carries, adjudicated on four witness confirmations across two publishers; and two apparent sanctoral gaps shown to be `Missae pro aliquibus locis` in the Missal's own words — particular feasts colitur is right to exclude, and which an overlay supplies. 761 tests, every year 1583-9999.
+
 ## [0.10.1] - 2026-08-21
 - **RG 33's third omission trigger**, the last unbuilt clause in the EF module. A II- or III-class vigil is omitted entirely not only when it falls on a Sunday or a I-class feast, but *"vel si festum cui praemittitur in alium diem transferri aut ad commemorationem reduci contingat"* — if the feast it precedes fails to keep its own day. Only the first two triggers were implemented; the code comment recording the third as unbuilt argued that no witness existed in the shipped data, and that was simply wrong: it fires on **1 744 days** across 1583–9999, in three shapes, every one classified against a pre-change binary. 1 199 are St Lawrence's vigil dropped because 10 August fell on a Sunday and reduced the feast to a commemoration; 545 are St John the Baptist's or Peter and Paul's dropped because the Sacred Heart or Corpus Christi took the feast's day and it transferred under RG 96. Both halves reduce to one question — is the feast the observed office tomorrow? — so the kernel asks it once per candidate after transfers settle, in a single pass with no fixed point: a vigil is a candidate only on its own day, never on its feast's, so suppressing it cannot change what the next day observes. `Precedence.rules` gains `vigil_feast`, which returns the slug of the feast a vigil precedes, because the kernel cannot infer it — only two of the five vigil/feast pairs share a slug stem. **Verified outside the project's own lineage, which is rare here**: O'Connell's *The Celebration of Mass* (1964) states all three triggers and independently confirms the five-vigil inventory and its class split, and the Latin Mass Society Ordo 2024–2025 gives 9 August 2025 as "OUR LADY on SATURDAY IV Cl W / Commem of St Romanus M" with no vigil — colitur's output field for field, and discriminating because that edition lists six other vigils including the Assumption's five days later. An earlier claim that the Ordo already in `docs/research/` corroborated this was **withdrawn**: it omits St Lawrence's vigil on every date it covers, including one where the rule says to keep it, so its agreement was a gap coinciding with the right answer. The clause's transferred-feast shape still has no day-level witness, which the allow-list note states rather than glosses. Two cited allow-list entries (C39 vs lectio, an adjudicated class vs the oracle); the RG 128 golden pin asserting St Lawrence's vigil is violet moved 2025 → 2027, its own comment having reasoned about the vigil's weekday and missed that the rule also looks at the feast's. 575 tests.