docs(citations): migrate register line-number pointers to stable anchors

68 "register line N" pointers in tracked code and data cited the
register's line numbers directly. The register has moved since: line 334
(cited six times as RG 91 entry 8) is now rose/colour prose; lines
375-376 (cited as "RG 109 leaves Advent/Lent Ember ordinary") are now
RG 26, the paragraph that refutes that exact reading; line 374 (cited as
RG 109(a)) is now RG 25's Latin; line 378 (RG 111) is blank.
expected-divergences.sexp even wrote "RG 91 line 332" three times, as
though the line offset were part of the rubric's own identity. Since
docs/research/rules-register.md is gitignored, no reader of the public
repo could resolve any of these, and -- since the register has moved on
its own copy too -- neither can the author's.

Replaced every pointer with an anchor that survives editing: the RG
paragraph/entry number itself (already present in nearly every case,
since the citation text already names "RG 91 entry N" or "RG 109(a)"
next to the stale line number -- the line number was redundant, not
load-bearing) plus a "§4" or quoted-subsection-heading pointer into the
register for readers who want to locate the fuller discussion. Also
corrected one genuinely stale content error found while migrating its
anchor: test_precedence_ef.ml's RG33 disposition-cases comment still said
"a I/II-class vigil impeded by any Sunday" -- the pre-Task-16
mistranscription the register itself corrects to II/III class; fixed the
prose alongside its citation, since leaving a wrong RG-class claim next
to a freshly-verified anchor would be worse than the stale line number it
replaced.

Covers lib/rites/rite_ef/precedence_ef.ml (18), precedence_ef.mli
(already fixed with item 1), test/test_precedence_ef.ml (44, including
two instances that only word-wrapped "register\nline N" across a line
break and so did not match a same-line grep), and
data/ef/expected-divergences.sexp (5, including three "RG 91 line 332"
instances). All 68 original pointers resolved -- none needed guessing;
every citation's target rule was already named in the surrounding prose.

No behaviour change: every edit is a comment/citation/data-note change.
Verified byte-identical `colitur day` output across 1583, 1900, 1902,
2008, 2011, 2026, 2038, 9999. 259/259 tests green.

1 files changed, 5 insertions, 5 deletions

diff --git a/data/ef/expected-divergences.sexp b/data/ef/expected-divergences.sexp
index 6b7bbdd..c8d863d 100644
--- a/data/ef/expected-divergences.sexp
+++ b/data/ef/expected-divergences.sexp
@@ -20,15 +20,15 @@
 ((id C1)
  (citation "RG 72-73 (Nativity/Epiphany season boundary, Jan 1-13) + RG 119a (white through \"expletum tempus Epiphaniae\"); register §3c item 1")
  (verdict colitur)
- (note "6-13 January is Christmastide in colitur, matching RG 72-73's explicit boundary at 13 January (corroborated independently by RG 119a's colour rule, register lines 241-246). lectio's efSeason switches to time-after-epiphany on 6 January. The season disagreement cascades into colour (green vs white) and, on 7-8 January specifically, into the ferial slug family name (colitur's ef-christmas-2-<weekday> vs lectio's ef-time-after-epiphany-1-<weekday> for the same two days).")
+ (note "6-13 January is Christmastide in colitur, matching RG 72-73's explicit boundary at 13 January (corroborated independently by RG 119a's colour rule, §3b). lectio's efSeason switches to time-after-epiphany on 6 January. The season disagreement cascades into colour (green vs white) and, on 7-8 January specifically, into the ferial slug family name (colitur's ef-christmas-2-<weekday> vs lectio's ef-time-after-epiphany-1-<weekday> for the same two days).")
  (expected_rows 368))
  ((id C2)
- (citation "RG 91 line 332 (\"Sundays of Advent, Lent, Passiontide, and Low Sunday\" -- I class, unqualified); extends register §3c item 4 (stated there for Advent only) to Lent on the same textual basis")
+ (citation "RG 91 entry 6, §4 (\"Sundays of Advent, Lent, Passiontide, and Low Sunday\" -- I class, unqualified); extends register §3c item 4 (stated there for Advent only) to Lent on the same textual basis")
  (verdict colitur)
- (note "Advent II & IV and Lent I-III are I class in colitur, matching RG 91's own unqualified line 332 (already confirmed for Passion Sunday, Palm Sunday, Low Sunday and Advent I, where lectio agrees). lectio's temporal_ef.go generic Sunday branch assigns Class2 and special-cases only Advent I, Passion/Palm/Low Sunday -- a lectio gap for the remaining five Sundays this entry covers.")
+ (note "Advent II & IV and Lent I-III are I class in colitur, matching RG 91's own unqualified entry 6 (already confirmed for Passion Sunday, Palm Sunday, Low Sunday and Advent I, where lectio agrees). lectio's temporal_ef.go generic Sunday branch assigns Class2 and special-cases only Advent I, Passion/Palm/Low Sunday -- a lectio gap for the remaining five Sundays this entry covers.")
  (expected_rows 215))
  ((id C3)
- (citation "RG 91 line 332 (same I-class Sunday rule as C2) + RG 131 (rose indult, Gaudete/Laetare, \"in Officio et Missa diei dominici tantum\")")
+ (citation "RG 91 entry 6, §4 (same I-class Sunday rule as C2) + RG 131 (rose indult, Gaudete/Laetare, \"in Officio et Missa diei dominici tantum\")")
  (verdict colitur)
  (note "Gaudete (Advent III) and Laetare (Lent IV): I class per C2's citation, plus the RG 131 rose indult. lectio has no Rose case in its EF colour function and zero rose rows in tridentine-calendar.ini -- Rose exists only on lectio's OF path -- so it prints violet for both Sundays.")
  (expected_rows 89))
@@ -58,7 +58,7 @@
  (note "Rogation Monday and Tuesday exist as colitur's own ef-rogation-{monday,tuesday} slugs (violet) whenever no higher-ranked saint intervenes. lectio computes no Rogation days at all and shows the plain paschaltide-week-6 feria instead.")
  (expected_rows 31))
  ((id C9)
- (citation "RG 95 (translation is a right of I-class FEASTS only) + RG 91 line 332 (Sundays are a separate table row, not feasts) + RG 109(a) (\"of a Sunday\" is a privileged commemoration, presupposing the Sunday stays put) + RG 96 (the general translation walk) -- the same reasoning already fixed for the analogous impeded-Sunday defect in Precedence_ef (Task 9)")
+ (citation "RG 95 (translation is a right of I-class FEASTS only) + RG 91 entry 6, §4 (Sundays are a separate table row, not feasts) + RG 109(a) (\"of a Sunday\" is a privileged commemoration, presupposing the Sunday stays put) + RG 96 (the general translation walk) -- the same reasoning already fixed for the analogous impeded-Sunday defect in Precedence_ef (Task 9)")
  (verdict colitur)
  (note "St Joseph (19 March, I class) colliding with a Lent Sunday (I class per C2): colitur keeps the Sunday observed and transfers Joseph to the next free day (usually 20 March; in years the walk is congested enough to cross Easter, as far as 1 April -- 2008, 2035, 2046, independently pinned by this project's own transfer tests). lectio instead puts Joseph ON the Sunday, displacing the Lenten Sunday office RG 95 says cannot be displaced by a translated feast.")
  (expected_rows 15))