fix(of): rename credo to creed, correct three comment inaccuracies

Fix round 1 for Task 3, addressing review findings.

Rename the exported creed function from `credo` to `creed` across the
.ml, .mli and test call sites, so the OF rite bundle can write
`creed = Rubrics_of.creed` with no gratuitous mismatch against
Rite.t's own field name, matching Rite_ef.Rubrics_ef.creed's own
naming. IGMR n. 67's own heading is "Professio fidei" and its body
says "Symbolum", never "Credo" -- documented as a comment on the
now-correctly-named function rather than kept as the exported name.
Local test helpers (credo_on_temporal, check_credo, test_credo_*) and
Alcotest description strings keep the Latin shorthand, since only the
call sites against RO.creed needed to change.

Corrected two comment inaccuracies:
- The Annunciation (25 March) is NOT always inside Lent, unlike St
  Joseph (19 March): 25 March sits after Easter's own earliest
  possible date (22 March), so in an early-Easter year (2008, Easter
  23 March) it falls inside the Easter Octave instead, or is
  transferred under Normae n. 60/5. No functional effect (gloria/creed
  read observed/temporal as actually resolved), but the stated reason
  was wrong.
- The Requiem-silence argument's counts: three All Souls Mass
  formularies (not two), and roughly 147 Gloria+Credo markers combined
  elsewhere in the Missal (not "over 150", which conflated Gloria
  alone with the combined figure) -- recounted directly against a
  fresh pdftotext -layout extraction rather than estimated.

Added: an independent corroborating citation for the Holy Thursday /
Easter Vigil Gloria exception, Vigil n. 31 -- the Easter Vigil's own
praenotanda ("VIGILIA PASCHALIS IN NOCTE SANCTA"), a THIRD
independently-numbered rubric set inside the same Missal PDF,
confirming the Vigil's own Gloria directly rather than resting solely
on the Holy Thursday rubric's forward-looking "usque ad" clause.

dune test: exit 0, 845 tests. make check (COLITUR_EXHAUSTIVE_SWEEP=1,
full 1583-9999 sweep): exit 0, 851 tests, 330.3s, both run unpiped in
the foreground. git diff --stat v1.0.0..HEAD -- lib/rites/rite_ef/
lib/kernel/ data/ef/ still empty.

1 files changed, 22 insertions, 15 deletions

diff --git a/lib/rites/rite_of/rubrics_of.mli b/lib/rites/rite_of/rubrics_of.mli
index 5325f12..ced9193 100644
--- a/lib/rites/rite_of/rubrics_of.mli
+++ b/lib/rites/rite_of/rubrics_of.mli
@@ -44,20 +44,27 @@ val gloria :
   bool
 
 (** Whether the Symbolum (Credo) is said, post-homily, at this day's Mass
-    (IGMR n. 67-68). Reads [temporal] for the identical reason [gloria]
-    does -- n. 68's own "diebus dominicis" is a plain, weekday-only fact,
-    unlike n. 53's Advent/Lent carve-out, so every Sunday qualifies
-    (Advent/Lent Sundays INCLUDED, unlike [gloria]) -- and [observed]'s
-    [rank] for n. 68's own "in sollemnitatibus", and [slug] against
-    {!requiem_slugs} for the identical reason [gloria] checks it. Narrower
-    than [gloria] on purpose: n. 68 has no "et festis" clause at all, so a
-    [Festum]-rank day (an ordinary feast) does NOT get the Credo -- the one
-    place this function's shape most visibly parts from EF's own RG
-    475(b)/(c). [date] is accepted, for signature symmetry with [gloria]/
-    [preface] and {!Rite.t}'s own shared shape, but genuinely unread: n.
-    67-68 carries no Easter-relative window of its own, unlike RG 475(d)'s
-    three EF octaves -- checked, not assumed (see the .ml's own citation). *)
-val credo :
+    (IGMR n. 67-68). Named [creed], not [credo]: IGMR n. 67's own heading is
+    "Professio fidei" and its body says "Symbolum", never "Credo" (checked
+    directly, igmr.txt:568-577) -- "Credo" is this file's own prose
+    shorthand, kept in comments and test labels, but the exported name
+    matches {!Colitur_kernel.Rite.t.creed}'s own field exactly, the same
+    "interface name matches the kernel field it satisfies" discipline
+    {!Rite_ef.Rubrics_ef.creed} already follows. Reads [temporal] for the
+    identical reason [gloria] does -- n. 68's own "diebus dominicis" is a
+    plain, weekday-only fact, unlike n. 53's Advent/Lent carve-out, so every
+    Sunday qualifies (Advent/Lent Sundays INCLUDED, unlike [gloria]) -- and
+    [observed]'s [rank] for n. 68's own "in sollemnitatibus", and [slug]
+    against {!requiem_slugs} for the identical reason [gloria] checks it.
+    Narrower than [gloria] on purpose: n. 68 has no "et festis" clause at
+    all, so a [Festum]-rank day (an ordinary feast) does NOT get the Creed
+    -- the one place this function's shape most visibly parts from EF's own
+    RG 475(b)/(c). [date] is accepted, for signature symmetry with
+    [gloria]/[preface] and {!Rite.t}'s own shared shape, but genuinely
+    unread: n. 67-68 carries no Easter-relative window of its own, unlike
+    RG 475(d)'s three EF octaves -- checked, not assumed (see the .ml's own
+    citation). *)
+val creed :
   temporal:(Vocab_of.season, Vocab_of.rank) Temporal.t ->
   observed:Vocab_of.rank Celebration.t ->
   date:Date.t ->
@@ -69,7 +76,7 @@ val credo :
     Missal's preface collection (over 50 individual texts across ~30 named
     occasions) against {!Preface.t}'s 15 EF-cited constructors, and why
     none of them is safe to reuse. Same three parameters as [gloria]/
-    [credo], all three genuinely unread -- kept only for {!Rite.t}'s shared
+    [creed], all three genuinely unread -- kept only for {!Rite.t}'s shared
     shape. *)
 val preface :
   temporal:(Vocab_of.season, Vocab_of.rank) Temporal.t ->